The questions buyers actually ask

A website built to be read and acted on by AI agents, not just browsed by people. Three things have to be true: the content is machine-readable rather than trapped in PDFs and image flyers; the site is wired to its own data, so it can recognise who is in front of it and route them to the right next step; and it exposes structured information an agent can actually use to answer a question about you. Most sites fail the first test on its own.

A conventional website assumes a human with a mouse. An agentic website assumes that some meaningful share of the traffic making decisions about you is software — an assistant researching suppliers, a shopping agent comparing options, a retrieval system deciding whether to cite you in an answer.

Three properties, in order of how badly most sites fail them.

1. Machine-readable content. Key information exists as structured text, not as PDFs, image flyers, or text baked into graphics. This is where nearly every site fails first, and it is not exotic — it is a spec sheet that only exists as a PDF, a seasonal offer that lives inside a JPEG, a service line described only in a video. To a retrieval system, none of that exists.

2. Wired to its own data. The site knows what it contains and can route on it — recognising intent, surfacing the relevant proof, moving a visitor to the right next step rather than a generic contact form. The Ask bar at the bottom of this site is a version of this: it reads the question, retrieves from actual content, answers, and routes.

3. Structured and exposed for agents. Clean schema markup, consistent entity data, and increasingly a machine interface. The emerging plumbing here is real: Microsoft's NLWeb turns a site into a conversational endpoint that also functions as a Model Context Protocol server, MCP has become the common standard for connecting AI systems to external tools and data, and on the commerce side the Agent Payments Protocol and the newer Universal Commerce Protocol are building the rails for agent-executed purchases.

What it is not. It is not a chatbot bolted onto a brochure. A widget that answers from a scripted FAQ is a support deflection tool. An agentic site is one where the underlying content and data are structured well enough that any agent — yours, or one you have never heard of — can use them.

Why it matters commercially. If the machine cannot read you, you do not enter the consideration set, and the consideration set is now formed before a human visits your site. Gartner finds that 75% of B2B buyers prefer a rep-free buying experience, and buyers spend only about 17% of the total purchase journey in contact with suppliers at all. Most of the decision happens where you are not in the room. An agentic website is how you are represented accurately while you are absent.

The honest sequencing. Almost nobody needs the advanced layer yet. Nearly everybody needs the first property. Get the content out of the PDFs, structure the pages, fix the schema — that alone moves you further than any agent integration, and it is the work that pays off regardless of which protocols win.

Sources: Gartner — the B2B buying journey, Model Context Protocol, Microsoft — introducing NLWeb

Mostly we do not say, and that is deliberate. Much of our best client work is protected by NDA, and companies preparing for a transition have real reasons not to appear in someone else's marketing. So we state outcomes without logos — industrial eCommerce grown from roughly $16,000 to more than $1M a year, a valve catalogue rebuilt to roughly 15,000 purchasable SKUs, a Tier II supplier whose two partners exited in just over four years with more than $20M each. Specifics in your category, and anything about who we work with alongside you, are for a conversation rather than a web page.

There is a version of this answer that sounds like an excuse, so here is the version that is actually true.

Confidentiality is standard in strategy work. But there are reasons specific to what we do. A meaningful share of our work sits close to a transaction: preparing a company for sale, strengthening a growth story before diligence, repositioning a business a private equity firm is about to take to market. Publicising that a company is working on its exit readiness is commercially harmful to that company. Even away from transactions, the work often involves telling a company that its promise is broken — and publishing that is not a case study, it is a liability for the client.

What we do instead: we state outcomes in results-first terms with real numbers and no logo, and we describe the shape of the work in your category — what usually breaks, what we usually find, what the sequence looks like — which is more useful than a logo anyway.

The trade we are making: we lose some credibility on first impression by not showing logos, and we keep the trust of clients who need discretion. Given that a large share of our best work happens in exactly those situations, that is not a close call. If you want to go further than this — references, or who else we work with in your space — ask, and we will talk about it properly.

We disclose, we separate, and you decide. We work across a small number of complex industries, so we occasionally serve companies that overlap at the category level. When that happens we say so before you engage, not after. Engagements are separated — no shared strategy, no competitive information moving between accounts. If we believe an overlap would genuinely compromise either client, we decline the work. When an engagement ends, the non-competition ends with it and we are free to work in that category again. What does not end is confidentiality: what we learned about your business stays inside your engagement permanently.

Conflicts are handled by disclosure rather than by a rule that would be easy to state and hard to keep.

We work across a small number of complex industries. Depth in those industries is the reason we are useful, and it means we occasionally serve companies that overlap at the category level. Pretending otherwise would be a claim we could not honour.

So the practice is this. Before an engagement begins we check our current book against your competitive set and tell you what we find. Where we work with companies in adjacent or overlapping spaces, the engagements are separated: different teams where it matters, no shared strategy, and no competitive information moving between accounts. If we believe an overlap would genuinely compromise either client, we decline the work. Where it would not, we tell you it exists and let you decide.

When the engagement ends, so does the restriction. We do not hold a category open indefinitely on the strength of work that finished two years ago, and we would rather say that plainly than let a prospect assume a protection that is not there. If exclusivity beyond the life of an engagement matters to you, raise it and we will talk about what is workable.

Two things that do not expire. Confidentiality is permanent — what we learned about your business stays inside your engagement, whether or not we are still working together, and whoever we work with next. And in transaction-related work we do not act for parties on opposite sides of the same process, and we do not carry information between them. Where we work with an advisory firm or sponsor across multiple engagements, each portfolio company's information stays inside that engagement.

Working with several brands inside one corporate group is not a conflict — it is one client relationship with multiple business units, and it is disclosed as such.

Yes. We built a coordinated multi-component programme for a client with a fixed launch window, then agreed to test one component in isolation first. The test produced genuinely useful data. It also stalled the wider rollout for months, because a phased test and a coordinated launch are different projects and we had not said so out loud. We had to go back to the client, reconcile scope and budget, and restart with a committed timeline. The lesson we kept: when a plan changes shape mid-flight, stop and rescope explicitly. Drift is not a decision.

We designed a programme with several components meant to launch together over a defined window. Partway in, we agreed with the client to test one component on its own before committing the rest. That was a reasonable decision on the evidence available, and the test worked — it produced real performance data we would not otherwise have had.

What it also did was break the coordinated launch. A programme where components reinforce each other and a sequential test-one-thing-at-a-time approach are two different projects with two different timelines and two different budgets. We slid from one into the other without ever naming the change. Months passed. The original launch window went by. Neither side was obstructing anything — the work was ongoing and mostly productive — but the thing we had actually agreed to do was no longer the thing we were doing, and nobody had said so.

How it got fixed. We wrote to the client, said plainly that the testing had been valuable and that the project had drifted, and asked for forty-five minutes to align on four things: what the testing had taught us, what remained, who owned each piece, and a committed launch date. We reconciled scope against budget in that meeting because the programme had genuinely evolved and the original numbers no longer described it.

What we own. We were closest to the work, we saw the shape change first, and we should have called it earlier. Waiting made the conversation harder and the correction more expensive. Not calling it was the failure — not the decision to test.

What changed in how we work. Any material change of shape now triggers an explicit rescope: what was agreed, what changed, what it now costs, what the new date is, in writing. It sounds bureaucratic. It costs one meeting and prevents four months.

If you want to check whether a firm will tell you when something is going sideways, ask them this question and see whether the answer has a date and a specific mistake in it.

That is not something we can answer here, and the reason matters. We do not sell tactics off a shelf like a typical marketing agency — we build a solution that stabilises the company and grows its value, and the shape of that is different for every business. Duration comes out of discovery, once we know what is actually broken and what it will take to fix. The one thing that is fixed is the front door: the Outside-In Analysis comes back about five business days after we have your domain, and it costs nothing. Start there and the rest becomes a real conversation instead of a guess.

Any firm that quotes you a duration before it has looked at your business is quoting you a template, and a template is the thing we are built to replace.

Engagements are scoped to what the Outside-In Analysis and Discovery actually find. What is broken in one company is a quoting process; in the next it is a promise the operation cannot keep; in the next it is that nobody can find them at the moment the decision gets made. Those are not the same job and they do not take the same time.

What we can tell you before any of that: the Outside-In Analysis takes about five business days, asks nothing of you but a domain, and costs nothing. It is deliberately the first step, because it turns every question like this one from a guess into something answerable.

Ignite XDS does not do the same thing for every client. Every engagement is unique, so this is worked out with you during discovery rather than answered from a web page.

Four phases of work, gated by three commitments. The work: Outside-In Analysis — we study how you show up to buyers, competitors, search and AI before you explain anything to us. Strategic Discovery — we pressure-test those findings with leadership, sales and operations to find the cause behind the symptoms. Priority Roadmap — what happens now, next and later. Execution Partnership — building the messaging, systems and infrastructure. The commitments that gate them are the Pre-Flight conversation, the Discovery Commitment, and Implementation. The phases are what we do; the commitments are what you agree to. Each step earns the next.

Most firms start Discovery by asking leadership to explain the business. We do not, and the reason is structural: you cannot get an unbiased read on a company from the people inside it, not because they are dishonest but because they are fluent. They have stopped seeing what a stranger sees in eight seconds.

Phase 1 — Outside-In Analysis. Before we ask you anything, we study what the market can already see: positioning and messaging, digital and AI visibility, competitive landscape, customer journey, digital experience, and proof and credibility. Then we bring it to you and ask one question — here is what the market appears to believe about your company; where are we right and where are we wrong? That question is the beginning of useful strategy, and it is impossible to ask without doing the work first.

Phase 2 — Strategic Discovery. The outside view shows symptoms. Discovery finds cause. We work with leadership, sales, marketing, operations, and customer-facing teams, looking for a specific list of failures: buyers who need too much explanation, sales processes that depend on two or three people, customer expectations operations cannot consistently meet, messaging that does not reflect what the company is actually best at, digital tools that create friction instead of confidence, and growth plans not tied to execution capacity.

Phase 3 — Priority Roadmap. Not a long document. A sequence. What creates the strongest business impact, and what should happen now, next, and later. Most strategy work dies here — it produces a deck nobody executes because it was never sequenced against capacity.

Phase 4 — Execution Partnership. Messaging, systems, digital infrastructure, campaigns, sales tools, customer experience. We become an extension of the team rather than a vendor delivering into it.

Running alongside those four phases are three commitments: the Pre-Flight conversation, where both sides decide whether this is worth starting; the Discovery Commitment, where you commit to the honest self-examination; and Implementation, where you commit to changing something. The phases are what we do. The commitments are what you agree to. We do not assume a long-term engagement before we understand the business, and we would rather lose the work at Pre-Flight than at month four.

What you receive, and when.

After the Outside-In Analysis: the document package and an offered walkthrough, yours to keep regardless of what happens next.

Discovery is its own engagement, and it is deliberately small — a focused, paid piece of work whose job is to pressure-test the opportunities the Analysis identified against your real numbers, your real team, and your real ambitions. It comes with a working-session agenda naming the outcomes the session has to reach, a pre-meeting information request covering only the questions that cannot be answered from outside, a live workbook that sits open on screen and is shared back within 24 hours as the official record, and — unusually — the framework of the final plan issued before Discovery begins, so you see the structure of the deliverable you will own before you commit to producing it. Nothing is finalised inside the meeting; either side can revise for 48 hours afterwards. Anything ranked uses the room average rather than the loudest voice, and every action item gets a named owner and a real date or it is not an action item.

At roadmap sign-off you receive the completed 12-and-24-month Operational Marketing Plan — the framework you saw before Discovery began, now filled in with your numbers, your owners, your dates, and your phasing — together with a work authorisation that states the scope and the cost. Scope changes after that are agreed in writing rather than absorbed quietly.

Through Execution we run a standing working session — weekly on active build work, less often once a program is steady — plus a separate monthly report-out. Those are deliberately two different meetings. The working session moves the work; the report-out looks at whether it is producing what we said it would, with the wider team in the room. The measures are agreed at the start of Execution so nobody is grading a moving target.

And you own it either way. If we finish Discovery and you decide not to continue, the plan is still yours. That is deliberate. A plan you cannot use without us is a hostage, not a deliverable.

No — nothing about what we are paid moves with what we find. There is no share of recovered margin and no success fee anywhere in how we work, which matters because it is the answer to a fair question: if we were paid more for finding money, we would find money whether or not it was there. We are not, so we can tell you plainly when there is nothing to find. What an engagement costs is a different question, and one we work out with you in discovery rather than answer here.

Two offices — Brighton, Michigan and Bradenton, Florida — and we work with clients nationwide from both. Location does not decide fit. Most of the work happens remotely and moves faster that way, but Discovery in particular benefits from time on site: walking the floor, sitting with the sales team, seeing where the customer experience actually happens. So we travel when it matters and stay remote when it does not. Hours are Monday to Friday, 9:00am to 5:00pm Eastern. Call or text 810-225-8350, or email info@ignitexds.com.

Where we are. Two offices, one team, clients nationwide.

Does location matter for the work? Less than it used to, with one real exception.

Most of what we do — analysis, positioning, content, systems, build work, weekly working sessions — runs remotely and often runs better that way. Remote sessions are shorter, better documented, and easier to schedule with people who are running a business at the same time.

The exception is Discovery, and it is a genuine one. Some of the most important findings are things you cannot get on a video call. Watching how a quote actually moves through the building. Sitting with the sales team and hearing what they say when the founder is not in the room. Standing in the dining room during a rush. Walking the plant floor and noticing that the thing everyone treats as normal is the constraint. The promise gap lives in the operation, and the operation has a location. So we travel when it matters.

Does location matter for who we take on? No. Nationwide, and the Michigan and Florida footprint has more to do with where the work already is — a dense industrial and manufacturing base in the Midwest, a hospitality and multi-location services market on the Gulf Coast — than with any geographic limit.

The AI visibility wrinkle, if you are a local or multi-location business. Local discovery has become city-by-city rather than brand-wide. Analysis of AI local visibility found assistants recommending only 1% to 11% of locations that perform well in conventional local search, and AI Overviews now appearing on a large majority of local searches. If you operate across several markets, your visibility is not one number — it is one number per city, and they diverge sharply.

Sources: Search Engine Land — AI local visibility report 2026

Yes — and the one I think about most is a case where the client did nothing wrong at all. We were deep into building an opportunity model and, in the final audit of our own work, found benchmarks we had used that the published evidence did not actually support, along with a cost estimate and a payback claim that were not properly grounded. We pulled all of it out before it reached the client. The model got smaller and the projected upside came down. It was also the first version of that model I would defend in front of a hostile analyst.

We were building an opportunity model for a client — the kind of analysis that puts a number on what a set of changes might be worth. Before delivery, we audit our own work: every figure has to trace to a source, and every source has to actually say what we claim it says.

Several did not. A benchmark we had leaned on turned out to be drawn from a different segment than the one we were applying it to. A cost estimate had no defensible basis. A payback claim rested on those two things and therefore on nothing. None of it was invented — every number had come from somewhere — but the chain from source to conclusion did not hold under examination.

We pulled all of it out. The model got smaller. The projected upside came down. Ranges replaced point forecasts, and the assumptions moved into a visible layer the client could edit and argue with, rather than sitting hidden inside the calculation.

Why we count this as a failure and not a success. We caught it — but we should not have needed to catch it, because it should not have been there. The pressure that produces this is real and worth naming: an opportunity model with a bigger number is more persuasive, and every incentive in this business pushes toward the bigger number. We had drifted toward it without deciding to.

What changed. Facts and models are now visibly separate in everything we produce. Observed evidence is stated and refuses to project. Anything modelled carries its assumptions, its capacity limits, and its ranges on the same page, and we name our two weakest assumptions out loud in the presentation before the client can find them.

What it means for you. Our numbers are smaller than they would otherwise be. They are also numbers we will defend line by line. If you are choosing between a firm quoting a large projected return and a firm quoting a smaller one with its assumptions attached, the second one is doing the harder thing.

Sometimes, and openly. Ignite XDS has its own team of long-tenured associates, each with a specific field of expertise. When a client needs a service or consulting outside that skill-set, we work with the client to select the right expert in that area — those specialists are typically managed by the Ignite team, though not always. What we will not do is quietly hand your work to whoever is available. Which specialists an engagement needs, if any, comes out of discovery.

The question underneath this one is usually about control and continuity: will the people who understand my business be the people doing the work.

The core work — the diagnosis, the positioning, the sequence, the decisions that shape the engagement — belongs to Ignite XDS and to the associates whose expertise the work calls for. Where something falls outside that skill-set, we do not pretend otherwise and we do not take the margin on work we are not the best at. We help you select the right expert in that area. Those specialists are typically managed by the Ignite team so you are not chasing them, though not always — sometimes the cleaner arrangement is a direct one, and we will say which we think it is.

That is a deliberately unglamorous answer, and it is the honest one. Our work is unusually context-dependent — we are looking for the gap between what a company promises and what it delivers, and that gap does not survive a handoff document. So the decisions stay with the people who found the gap, and the specialist work sits around them.

Which specialists, and whether any are needed at all, depends entirely on what the roadmap turns out to require. Ignite XDS does not do the same thing for every client. Every engagement is unique, so that gets worked out with you during discovery.

Marketing that starts from whether the organisation can actually deliver on the brand promise, rather than from campaigns. It treats operations, logistics, production, and service as part of the marketing system, because those are the functions that keep or break what the marketing said. That is why the questions sound different — clients tell us no marketing person has ever asked them this. The goal is improved EBITDA and a higher valuation, not clicks or impressions.

It costs nothing, and we need nothing from you. No data, no NDA, no interviews, no access to your systems — the entire analysis is built from what is already public, which is exactly what makes it honest. Turnaround is about five business days from the time we have your domain. The documents arrive first and are written to stand on their own; if the findings are useful, the next step is a 45-minute conversation to walk through them together. No obligation, no pitch, no proposal in the room. It is a five-figure piece of work, it is yours to keep either way, and there is nothing to buy in it.

The cost is zero, and that deserves an explanation rather than an exclamation mark.

Most engagements begin by billing you to explain your own business — weeks of interviews and a synthesis deck, much of it spent constructing a picture we could have built independently, except now you have paid for it and it is shaped by internal assumptions rather than outside reality.

Doing the analysis first inverts that. We arrive with a grounded read your leadership can validate or correct in a single session. That eliminates false assumptions early and makes everything after it shorter and sharper. We are not being generous. We are removing the least productive phase of a normal engagement and absorbing the cost, because it makes the rest better.

What we need from you: your website. That is it. No data pull, no internal scramble, no NDA, no conversation with your team. We do it the way a buyer or a competitor would — from the outside, with no access — because that is the only view that is not already contaminated by what you believe about yourself.

What that means practically: your competitors could run the same analysis on you tomorrow. Everything in it was already visible to anyone willing to look. The uncomfortable version of that sentence is that some of them probably have.

Turnaround is about five business days from the time we have your domain. The fieldwork itself is roughly two days of concentrated work; the rest is assembly and review, because we do not send anything we have not checked.

The documents arrive first, and they are written to stand on their own. If the findings are useful, the next step is a 45-minute conversation to walk through them together — offered, not required. No obligation, no pitch, and no proposal in the room. Plenty of people read the package and never take the call, and that is a legitimate outcome.

What you keep: all of it, in full, whether or not we work together. No expiry, no strings, no requirement to take a second meeting. If the findings are useful and you want help, there is a conversation. If not, you have a clear read on how the market sees you, which is worth having either way.

Six documents, and we produce them before we ever speak to you. A Landscape Brief on how the market actually sees you versus how you describe yourself. An AI-Era Visibility Report testing a dozen real buyer searches to see who gets named and what the absence is worth. A Persona Snapshot of the people who decide your revenue and what each needs to see to trust you. A Persona Resonance report scoring how your site actually lands with each of them. An Opportunity and Action Brief with the highest-value moves and the risk of leaving each alone. And a cover letter that tells you plainly there is nothing to buy in it.

We don't ask you to tell us about your company. We go and find out.

The analysis is built entirely from public information — no interviews, no data request, no NDA, no access to your systems. That is not a limitation, it is the point. Everything in it is what a buyer, a competitor, or an acquirer sees when they look at you, which means it is the only honest starting position available.

What you receive.

The Cover Letter. What we did, over what period, and what is enclosed. It also says the thing most firms leave out: there is nothing to buy in any of it. If the findings are useful, the next step is a conversation.

The Landscape Brief. How the market sees you today against how you describe yourself. Your company at a glance, drawn from what is publicly visible. Signals of momentum set beside signals of plateau — and specifically, the contradictions worth sitting with. Where competitors feel stronger to a first-time buyer, and where you should be winning and are not. It closes with a section written directly to leadership, because the CEO, the CFO, and the person running sales are each looking at a different problem.

The AI-Era Visibility Report. A dozen realistic buyer searches — the actual sentences your customers type, not keyword-tool abstractions — run against live answer engines. Who gets named, whether your own site is ever cited, which competitors take the slots you should hold. Then what the absence is worth, with the assumptions stated so you can argue with them. Most companies discover here that they are missing from queries their own homepage makes a claim about.

The Persona Snapshot. The people who actually decide your revenue, usually five or six of them. For each: what they are looking for, what they secretly fear, what they need to see online before they will trust you, how they genuinely research, and the exact moment they decide you are not for them. That last one is the useful part, and it is almost never on a slide.

The Persona Resonance Report. We walk your site the way each persona would walk it, in the order they would arrive, asking only the questions they came to ask — then score how it lands, out of five, with the reasoning. Site-wide issues that damage everyone get their own section. This is where the specific, uncomfortable, entirely fixable things surface.

The Opportunity and Action Brief. The verdict, then the five highest-value moves. Each carries its evidence, its business impact, the risk of inaction, and a time horizon. It ends with what must be true to capture them, and the choice stated plainly.

Where there is a portfolio, a seventh: a Corporate Appendix placing the business inside the wider structure — how concepts, brands, or units relate, where the parent story is strongest, and where it is doing the least work.

What makes it different from a free audit. A free audit is generated by a tool in ninety seconds to capture an email address. This is analyst work: real queries run against live engines, named competitors compared line by line, personas built from how your buyers actually behave, and a verdict someone is prepared to defend in a room. It is a five-figure piece of work, it costs you nothing, and it is yours to keep whether or not we work together.

A premium beverage brand built its entire promise on a genuinely distinctive production story — the kind of detail that wins someone over completely, once they hear it. On a shelf, next to forty other bottles, there is no one to hear it, and the price varied depending on which channel you bought through. The brand was not failing to be interesting. It was making a promise its distribution channel was structurally incapable of carrying.

Some promise gaps are operational. This one was architectural.

The brand had a real story — a production detail distinctive enough that people who heard it remembered it and told other people. In a tasting, in a conversation, behind a bar, it converted. The founders knew this, because they had watched it work in person hundreds of times.

The strategy was therefore to build awareness so more people would hear it. Reasonable. Wrong.

The channel could not carry the story. A bottle on a shelf has a few square inches and about two seconds. Nobody is standing beside it to explain. Every dollar of consumer awareness spend was pushing people toward a moment of decision at which the actual differentiator was structurally unavailable. The promise was being made in one place and broken in another, and no amount of additional volume fixed that — it just increased the number of people who encountered the brand at its least persuasive.

Two more gaps sat underneath it. Price varied by channel, so a customer who found the brand in one place and looked for it in another got a different number, which quietly undermines a premium position. And a quality claim central to the brand's credibility had never been independently verified — which is fine right up until an informed buyer or a category buyer asks, at which point an unverified claim is worse than no claim.

What changed, and the order mattered. Credibility and pricing before demand generation: verify the quality claim properly, set one controlled price per SKU across every channel, and settle the positioning. Then shift the primary channel to where the story can actually be told — on-premise, with the people who hand the product to the customer and can say the sentence that makes it land. Then, and only then, build broader visibility on top of a foundation that holds.

The general lesson. Before you spend on awareness, check that the place where the decision gets made can carry the thing that makes you worth choosing. If it cannot, awareness spending accelerates the wrong outcome.

A distributor with thousands of catalogue pages promised fast, expert quoting and delivered exactly that — until the quote was sent, at which point it entered a filing cabinet. There was no system for following up, so the difference between a won quote and a lost one was whether a busy person happened to remember it. Pages that generated a quote request converted at roughly nine times the rate of pages that did not, which meant the highest-value moment in the entire business was the one with the least process behind it.

The promise was competence: send us a requirement, we will quote it accurately and quickly. The company kept that promise. Quotes went out fast and they were good.

Then nothing.

There was no CRM. A quote left the building and its fate depended on whether someone remembered to chase it, which meant follow-up correlated with how busy that week had been rather than how good the opportunity was. Nobody could answer basic questions: what proportion of quotes close, why the lost ones were lost, how long the first response actually takes, which product families win and which do not.

When we analysed several months of behaviour against sales, the shape of the problem became specific. Product pages that produced a quote request converted at roughly nine times the rate of pages that did not. Organic search supplied about half of all quote requests. The quote was, unambiguously, the highest-leverage event in the business — and it was the only significant commercial process with no system attached to it.

The promise gap here is subtle, because on the surface everything is working. Marketing generates traffic. The catalogue converts. Sales quotes accurately. Each function is performing. The promise breaks in the space between functions, which is exactly where nobody is looking, because no individual scorecard covers it.

What changed. A quote-to-close pipeline with a defined follow-up cadence, mandatory closed-lost reasons, and dashboards for close rate, first-touch speed, and win rate by product family and customer class. Not sophisticated. Just present.

The other finding worth stating, because it recurs across distribution: the product pages all ran from a single shared template fed by catalogue data. Which means the fix was never page-by-page redesign — it was the template and the completeness of the underlying product data. In a business with thousands of SKUs, product data quality is the customer experience. It is also what determines whether an AI system can understand and recommend you, so the same work pays twice.

A single-location restaurant was winning the visibility battle — AI assistants recommended it for most local queries and ranked it first for several, ahead of far more marketing-sophisticated competitors. It was still losing guests, because the questions people actually ask before choosing where to eat had no published answers: prices, kids' menu, gluten-free handling, late hours. It was already found, and not yet chosen.

This engagement inverted what everyone expected, including us.

The assumption going in was a visibility problem. Single location, modest marketing operation, competing against a market full of better-resourced restaurants. The obvious diagnosis was that nobody could find it.

The opposite was true. When we tested a set of real local queries against live answer engines, the restaurant was cited on the large majority and ranked first on several — outperforming competitors with far more marketing sophistication. Visibility was not the constraint. It was arguably an asset.

What it lost were the questions that decide the meal. Not "where should I eat" — it won those. It lost "how much is that", "do you have a kids' menu and what does it cost", "can you handle gluten-free", "are you open late". Every one of those is a question a real guest asks with their thumb, in a car, deciding between two options. And every one had no published answer, so the machine could not give one, so the guest went to the restaurant that had told them.

The scale of that was visible in the search data: hundreds of branded menu searches every month — people specifically looking for this restaurant's menu, by name, and not finding the facts they needed. That is not a marketing failure. That is demand you have already earned, arriving at a closed door.

The pricing incoherence made it worse. A half portion was priced against a full portion in a way that implied the smaller size cost substantially more per pound. Not a strategy — an artefact of prices set at different times by different people. But a guest comparing the two sees it instantly, and what they conclude is not "inconsistent pricing". It is "this place is expensive".

What changed. Publish the facts first — prices, hours, dietary handling, portion definitions — in a form both people and machines can read, before spending anything on further visibility. Then correct the price ladder so the menu tells one story. Then build out the channels where the real headroom sat, which turned out to be catering and off-premise rather than more covers.

The general lesson for multi-unit operators. Visibility and conversion are separate problems and they fail independently. Being recommended is worth nothing if the answer stops one question short of the decision. Most restaurant marketing spends on the first problem while the second one quietly eats the return.

In a sale process, the growth section is the most expensive page in the book, and it is the one nobody rewrites. Historical financials get verified in diligence; the growth narrative does not, so buyers do the only rational thing available and discount it. The promise gap in a transaction is between the story a seller tells and the story a buyer can underwrite — and every figure that cannot be traced to verified financials makes the figures that can look less trustworthy too.

Here is the mechanic. Historical financials get verified in diligence — a quality of earnings analysis tests whether reported earnings are real, recurring, and transferable. What does not get verified the same way is the growth narrative: the story about why the next five years justify the multiple. So buyers do the only rational thing available to them. They discount it.

The promise gap in a transaction is between the story the seller tells and the story a buyer can underwrite. They are not the same document. A seller writes to impress. A buyer reads to find the parts they can defend to an investment committee, and mentally deletes the rest.

Three problems are typical of the genre. Growth projections that are not traceable to verified financials, so an analyst cannot reconcile them. A most-optimistic scenario scaled so aggressively that it anchors the whole document — and when one number looks unserious, it contaminates the credibility of the numbers beside it. And small internal inconsistencies: headcount and descriptive details that do not match between sections. Individually trivial. Collectively, they are exactly what a diligence team notices, and what they conclude is that nobody checked.

What fixes it. Rebuild every figure from independently confirmed financials rather than internal estimates. Scale back the aggressive scenario deliberately — not to be modest, but because an anchor a buyer cannot defend costs more than it gains. Present scenarios as ranges with their assumptions attached, so a buyer can argue with the inputs rather than dismiss the output. And reconcile the document against itself so every detail matches everywhere it appears.

The resulting narrative is less exciting and considerably more valuable, because it survives contact with an analyst.

Why this is a promise gap and not just document quality. The promise a seller makes is: this business will keep growing after you own it. The experience a buyer has is reading a growth section they cannot verify. The gap between those two is priced — quietly, in the multiple, without anyone naming it in a meeting. Businesses with documented, repeatable revenue generation trade at meaningfully higher multiples than equivalent businesses that cannot evidence it. Being able to prove how customers are won is not a marketing asset. It is a valuation input.

The general lesson. Fix the growth story before the process starts, not during it. Once a buyer has discounted the narrative, you are negotiating against their number rather than presenting yours.

A precision manufacturer promised responsive quoting and invited RFQs from every page of its website. The form had been silently failing — submissions were being blocked before they ever reached the sales inbox, and nobody knew, because a form that fails quietly looks exactly like a market that is not interested. Fixing the delivery path took an afternoon. The year of quotes that never arrived did not come back.

This one is uncomfortable because it is so ordinary.

The company was good. Real capability, real engineering depth, a legitimate claim to being responsive. The website said so, and a Request a Quote form sat on every relevant page. The form was broken. Not visibly — the on-page confirmation still appeared, so a visitor filling it in believed they had reached someone. Submissions were being blocked in transit and never reached the sales inbox.

From inside the business, this is invisible. It does not present as a failure. It presents as a quiet quarter, a soft market, a website that does not really generate anything.

That is the promise gap in its purest form. The promise was: send us your requirement and we will respond. The experience was silence. Nothing about the marketing was wrong. Nothing about the sales team was wrong. A delivery mechanism nobody had tested since installation was breaking the promise on every submission.

The second gap on the same engagement was the opposite problem, and it is more common than people admit: the promise was too big. The site claimed capabilities the company did not actually offer, certifications it did not hold, and response commitments it could not consistently meet. Every one had been added in good faith by someone trying to look competitive. Together they created a different kind of gap — a buyer arriving on the strength of a claim the operation could not honour.

What changed. We tested the submission path end to end and rebuilt it, then verified delivery with a live submission rather than assuming. Then we cut the claims back to what was demonstrably true and rebuilt the positioning around what the company genuinely leads with. The site got shorter and more specific. It also became defensible, which matters more than it sounds — an overstated capability does not just fail to convert, it produces the wrong enquiries and burns the sales team's time disqualifying them.

The general lesson. Before you spend anything on generating more demand, verify that the demand you already generate arrives. We check this on every engagement now, and we find something broken more often than we would like.

A commercial roofing contractor working on hospitals, plants and large institutional buildings had a website that read like every residential roofer in the region. It also offered a free multi-building walkthrough to anyone who asked, which meant its most senior expertise was being given away to unqualified prospects while serious buyers had no way to tell it apart from a company that reroofs houses. The fix was not more marketing — it was making the difficulty of the work visible, and turning the free walkthrough into a paid diagnostic with a real deliverable.

The promise was: we handle roofs that most contractors cannot. Complex access, occupied buildings, operating conditions where failure is expensive, systems that require genuine specialisation.

The experience the market got was a generic local roofing company. Same imagery, same language, same structure as any residential contractor. A facilities director responsible for a hospital roof, comparing three websites, had no way to distinguish thirty years of complex commercial work from a firm that had never worked on an occupied building. The differentiation was real. It was simply not legible.

The second gap was in the offer itself. The company offered a free assessment — walk the buildings, look at everything, come back with findings. In practice that meant the most experienced person in the business spending a day on a roof for anyone who asked, with no qualification and no commitment. Meanwhile the word "assessment" was doing two jobs at once, sometimes meaning a diagnostic visit and sometimes meaning actual work, which left prospects unclear what they were agreeing to.

The third gap was proof they could not substantiate. Several credibility claims on the site could not be evidenced when we asked for the backup. Some were probably true and simply undocumented. Some had quietly stopped being true — a certification programme that had been discontinued, a tenure statistic nobody could source. Every one of those is a liability the moment a serious buyer checks.

What changed. The qualification logic was rebuilt around what actually makes the work hard — operating conditions, access difficulty, system complexity, and the consequences of failure — rather than square footage. Square footage does not distinguish a warehouse from an operating theatre. Difficulty does.

The free assessment became a paid Inspection: a defined diagnostic product with a mapped roof, pinned findings, condition grading, remaining-life framing, and a written repair proposal. That single change fixed three things at once. It qualified prospects, because people who will not pay for a diagnostic were never going to buy the work. It gave the expertise a price, which is what makes buyers value it. And it separated diagnosis from work, so both sides knew what had been agreed.

Unsupported claims were removed or flagged as things still owed rather than published on faith.

The general lesson for trades and service businesses. If your differentiator is that the hard jobs are the ones you do well, your marketing has to make the difficulty legible. Buyers cannot infer competence from photographs of finished work — every contractor has those. They can infer it from how precisely you describe the conditions you work in.

A few situations, and we would rather say so early. If you want a single tactic executed — run these ads, build this page — you do not need us and we will be expensive for what you are asking. If the business is pre-revenue or still searching for what it sells, the work assumes there is something to align. If leadership is not prepared to hear that the problem may sit in operations rather than marketing, the Outside-In Analysis will be uncomfortable and the engagement will stall. And if the decision is purely about price, someone will always be cheaper.

This is the most useful answer on the site, because a firm that will not name its own limits is telling you it will take anything.

You want one tactic executed. If you know exactly what you need — a campaign run, a page built, a channel managed — hire a specialist. They will do it well and cost less. We are built to find out what is actually wrong, which is wasted money if you already know.

The business is pre-revenue or still finding its product. The whole method assumes there is a promise being made and an experience being delivered, and that the gap between them can be measured. Before product-market fit there is nothing stable enough to align. Different problem, different kind of help.

Leadership is not prepared to hear that the problem is not marketing. This is the real disqualifier and it has nothing to do with size or sector. The Outside-In Analysis frequently concludes that the constraint is a quoting process, a fulfilment failure, or a promise the operation cannot keep. If the expectation is more leads and no internal change, the engagement stalls at exactly the moment it becomes valuable. Bain's finding that 80% of companies believe they deliver a superior experience while only 8% of their customers agree is not a statistic about other people. Everyone is in the 80% until someone shows them otherwise.

The decision is purely about price. Someone will always be cheaper, frequently by using offshore labour or by selling you tactics. If lowest cost is the deciding factor, we will lose and we should.

You need speed above everything. If the timeline is two weeks to launch, there is no room for the work that makes it worth launching.

A note on how we handle it. When it is not a fit we say so at the Pre-Flight conversation, before anyone has spent money. Occasionally the Outside-In Analysis itself concludes the gap is small and you do not need us. We have written the analysis to make that possible — it costs us a few days and saves both parties an engagement neither should have started.

Sources: Bain — Closing the Delivery Gap

Every company makes a promise. Marketing and sales make it; operations, logistics, production, and service either keep it or quietly break it. When a company is small the founder holds the whole promise together by hand. As it grows it fractures across departments that no longer share the same plan, and nobody is looking at the whole. A concrete version: the website promises engineering partnership, the quote takes eleven days, and the buyer has already chosen someone else. Nothing in that sequence is a marketing problem, but marketing gets blamed for it. That gap is where revenue, margin, and enterprise value leak out.

This is the organising idea behind everything we do, so it is worth stating precisely rather than warmly.

The promise. Your marketing and sales make a claim to the market — explicit in the words on the site, implicit in the design, the response time, the confidence of the proposal. We are the engineering partner who solves hard problems. We are the operator who never misses a delivery date. We are the brand worth paying more for.

The experience. Operations, logistics, production, and service either keep that claim or quietly break it. Not dramatically. Quietly. The quote takes eleven days. The spec sheet only exists as a PDF. The person who knows the answer is on holiday. The second order arrives configured differently from the first.

The gap is where value leaks. Not just revenue — margin, because you discount to compensate. Not just margin — enterprise value, because a buyer paying a multiple is buying the reliability of the promise, and an inconsistent one gets discounted at diligence.

Why it opens as you grow. When a company is small the founder holds the promise together personally. They quote, they call, they catch the problem before the customer sees it. That does not scale, and the failure is not visible on any dashboard. The promise fractures across departments that each optimise their own metric, and no single person is looking at the whole. Every function can be performing well while the promise is broken.

The evidence this is systemic. Bain surveyed 362 firms and found 80% believed they delivered a superior experience to their customers. 8% of those customers agreed. The same study found more than 95% of management teams describe themselves as customer-focused, but only 30% organise the functions of the company to deliver a superior customer experience, and only 30% maintain effective customer feedback loops. That is not a customer service problem. It is a structural one: companies believe their own promise because nothing in how they are organised would tell them otherwise.

A concrete example. An industrial manufacturer positions itself as the responsive engineering partner. The site says it, the sales team means it, the engineers are genuinely excellent. A buyer sends an RFQ. It sits for eleven days because quoting requires a specific person who is also running production. By day four the buyer has three quotes from competitors who are worse engineers and faster responders. The company loses the work and concludes the market is price-driven. It is not. The promise was broken by a process nobody thought of as marketing.

Fixing that is not a campaign. It is not a rebrand. It is finding where the promise breaks and rebuilding the part of the business that breaks it.

Why we start outside. You cannot see your own gap. You are fluent in your business — you have stopped noticing what a stranger notices in eight seconds. That is why the first thing we do is look at you from outside, before you explain anything.

Sources: Bain — Closing the Delivery Gap

Real numbers, without the logos. Industrial eCommerce grown from roughly $16,000 to more than $1M a year across a platform spanning 16 websites, 11 languages, and over 100,000 SKUs. A valve catalogue rebuilt into roughly 15,000 purchasable SKUs so distributors could actually search and buy. A Tier II automotive supplier with stale positioning and flat growth whose two partners exited in just over four years with more than $20M each. A consumer brand that moved off the one-in-four repeat-purchase benchmark. What none of that means is a number we will promise you.

What we have actually done.

A global controls company with a fragmented digital presence and an underdeveloped eCommerce channel. We supported a global platform across 16 websites, 11 languages, and more than 100,000 SKUs. Annual eCommerce revenue went from roughly $16,000 to more than $1M.

An industrial valve company whose catalogue was too complex for distributors and end users to search, compare, and buy from. A limited online product experience became a structured digital catalogue with roughly 15,000 purchasable SKUs.

An industrial automation manufacturer's rep that grew from an 8,000 sq ft facility to 70,000 sq ft, supported with brand, website, customer portals, inventory systems, RMA systems, inspection tracking, strategic planning, and ongoing execution.

A privately held Tier II automotive supplier with strong capability, stale relationships, unclear positioning, and plateaued growth. We clarified the advantage, rebuilt the brand, launched a stronger website, equipped sales and engineering, and supported go-to-market. The two partners exited in just over four years with more than $20M each.

A dealership with an established community presence that needed a clearer growth plan and better use of its database. Brand updates, physical experience improvements, BDC support, segmented outreach, and ongoing sales reviews. Sales went from roughly 200 vehicles a month to more than 400.

A consumer products brand winning first orders and losing most buyers before a second purchase. We sharpened the brand story, rebuilt the product and comparison experience, and stood up a segmented repeat-purchase engine. Retention moved off the roughly one-in-four D2C benchmark, and repeat revenue became the growth driver instead of constant new-customer acquisition.

How to read those numbers, and how not to. Every one of them took years and involved a client doing hard things. We were part of the reason each happened; we were not the only reason, and we will not claim otherwise. Marketing that presents itself as the sole cause of a business outcome is misrepresenting how businesses work.

What we will not do is promise you a number. Not a percentage, not a multiple, not a timeline to ROI. Anyone who does is either guessing or selling. What we will tell you, after the Outside-In Analysis, is where we think the gap is and roughly what it appears to be costing — and that estimate comes with its assumptions attached so you can argue with them.

What outcomes we actually optimise for. Not clicks, not impressions, not lead volume. Improved EBITDA and a higher valuation. Those are slower to move and harder to attribute, which is precisely why most firms measure the other things.

SEO gets you ranked on a page of results. AEO — answer engine optimisation — gets you into the answer sitting above those results: AI Overviews, featured snippets, voice responses. GEO — generative engine optimisation — is the wider practice of structuring your content and presence so large language models retrieve, summarise, and cite you accurately. What changes in practice: SEO competes for clicks, AEO and GEO compete for inclusion. With 68% of Google searches now ending without a click, ranking and being found have stopped being the same thing.

The one-sentence versions. SEO — search engine optimisation. Getting your pages to rank in a list of links. AEO — answer engine optimisation. Getting your content selected for the answer that resolves the query before anyone clicks: AI Overviews, featured snippets, voice results. GEO — generative engine optimisation. Structuring content and presence so generative systems retrieve, summarise, and cite you accurately across ChatGPT, Gemini, Claude, Perplexity, and AI Overviews.

The three overlap heavily and the terminology is not settled — GEO is also called AIO and AI search optimisation, and plenty of people use AEO and GEO interchangeably. Arguing about the labels is a waste of a meeting.

What actually changes in practice. SEO competes for the click; AEO and GEO compete for inclusion. Under SEO, position one was the prize because position one got the traffic. Under AEO and GEO, the machine returns two to five options and that is the consideration set. If you are not among them, your reputation, your reviews, and your thirty-year history are not part of the decision, because the decision happened before anyone reached a website.

The unit of optimisation shrinks. SEO optimises a page. AEO and GEO optimise a passage — a self-contained block of text that answers one question completely enough to be lifted out and attributed. A page that is excellent as a whole but has no cleanly extractable answer in it can rank and still never be cited.

Machine-readability replaces keyword density as the binding constraint. Information trapped in PDFs, offers living inside image flyers, service lines buried behind JavaScript — none of it exists to a retrieval system. The most common cause of invisibility we find is not weak content. It is good content in an unreadable container.

The numbers behind the shift. 68.01% of US Google searches ended without a click in the first four months of 2026, up from 60.45% in 2024 and 49% in 2019. Where an AI Overview is present, click-through to the top organic result falls 58%, measured across 300,000 keywords. The compensating finding: visitors who do arrive from AI search convert at roughly 4.4 times the rate of conventional organic. Less traffic, better traffic.

The honest caveat, which most firms selling this will not give you. The widely quoted "GEO improves visibility by 40%" is a ceiling from one metric in one experimental configuration, where the source document had already been handed to the model. A July 2026 survey reviewing 45 studies across the field placed that claim in its lowest confidence tier and found no technique with a stable, cross-platform effect on discoverability. The direction of travel is not in doubt. The magnitude of any specific tactic is. Anyone quoting you a guaranteed AI visibility percentage is selling certainty that does not exist yet.

So what should a company actually do? Make the content machine-readable. Answer real questions in self-contained passages. Get the structured data right. Publish the facts that only you can state. Then measure whether you appear in the answers your buyers get — which is what the Outside-In Analysis does.

Sources: SparkToro — zero-click search, 2026, Ahrefs — AI Overviews reduce clicks, Semrush — AEO vs SEO, Aggarwal et al. — the founding GEO paper, Survey of 45 GEO studies

It is the first thing we look for, before we ask you to spend anything: money the business has already earned but is not collecting. Pricing that drifted below what the market carries. Settings inside your own systems that quietly leak margin. Demand you generate and lose. Customers who bought once and were never contacted again. When we find it, the growth work can be funded from recovered margin rather than new budget, and the risky part of the program comes last. It is not always there — some businesses are priced correctly and run tightly. But we look every time, and we will tell you plainly when the answer is no.

Every growth conversation has a second conversation running underneath it: this sounds right, I don't have the budget for it, and I'm not sure I can justify it to my partner, my board, or my banker.

The Wedge Layer is how we answer that, and it is a search we run during Discovery rather than a service we sell. The principle is: find the money first, prove it can be moved, then scale.

What we are looking for. Four places, in the order they usually pay.

1. Pricing that has drifted. The most common and the least examined. Prices set years ago, discounts that quietly became defaults, a menu or rate card that has not moved while input costs have. The leverage is unlike anything else available to a business: McKinsey's analysis of S&P 1500 economics found a 1% price improvement, with volume flat, produces roughly an 8% increase in operating profit. Nothing in marketing competes with that arithmetic.

2. Settings inside your own systems. Not strategy — configuration. An upcharge set to the wrong value in a point-of-sale system. A shipping rule nobody revisited. A discount code with no expiry. Duplicate items that split reporting so nothing looks important enough to fix. These are invisible from a P&L, because they never appear as a line. They appear as a slightly lower margin that everyone has learned to accept.

3. Demand you already generate and lose. Enquiries that go unanswered, a website converting at a fraction of what it should, a quote process that outlasts the buyer's patience, a form that has been silently failing. This is the cheapest revenue in any business, because you have already paid to create it.

4. Relationships you already own. Customers who bought, were satisfied, and were never contacted again. Accounts served in one location but not the other three. Warm lists that worked and quietly stopped. In most businesses this is the single largest pool, and it is almost never worked, because nobody's job description covers it.

Why we sequence it first. Two reasons, and only one is generous.

The generous one: funding growth out of recovered margin means the work does not compete with capital expenditure or hiring, which is what usually kills it at board level. It also inverts the risk. The cheapest, lowest-risk move goes first and is measurable in weeks. If it proves out, everything after it is not a bet — it is reinvestment of money the engagement itself produced.

The self-interested one: a wedge finding is early, concrete proof that we are looking at the whole business rather than the marketing corner of it. It is also the fastest way to demonstrate that the gap between what a company promises and what it delivers is real, quantifiable, and sitting in the data.

The part most firms would leave out. The wedge is not always available. Some businesses are priced correctly. Some have tight operations and no leak worth naming. Some are demand-constrained rather than conversion-constrained, which means there is no captured demand to recover because there was not enough demand to begin with — a genuinely different problem requiring a genuinely different answer. And some owner-led businesses have economics where a self-funding claim would be misleading, so we do not make one.

When that is the case we say so, and the growth investment is simply an investment. We have written diagnostics that deliberately carry no self-funding claim at all, because the evidence did not support one. We look every time. We do not always find. We never manufacture.

Sources: McKinsey — The Power of Pricing

Mitch Lipon owns and runs Ignite XDS, and works with a team of long-tenured associates, each with their own specific field of expertise. The people who sell the work do the work — there is no pitch team that disappears after signature. Which of those associates is on your engagement depends on what the work turns out to need, because Ignite XDS does not do the same thing for every client. That gets settled with you during discovery rather than published as a chart.

This question is usually asking two things: are you buying a person or a bench, and will the people in the pitch be the people in the work.

Mitch Lipon owns and runs the firm and is in the room for strategy. Around him is a team of long-tenured associates, each with a specific field of expertise rather than a general one. Long-tenured matters more than it sounds: the expensive thing in work that spans positioning, operations and systems is re-explaining your business, and you should have to do that exactly once.

The most common way a strategy engagement fails is not bad strategy — it is the senior person disengaging after the sale. Industry research on why client relationships end puts delivery dissatisfaction as the number one reason clients leave, while agencies rank it seventh. The gap between those two rankings is the whole problem: agencies do not know they are failing, because the person who would have noticed left the account.

Who is involved in a particular engagement, and at what cadence, follows from what the roadmap requires. Ignite XDS does not do the same thing for every client and every engagement is unique, so that is worked out with you during discovery rather than answered from a web page.

Privately held companies that have already built something valuable and know the next stage will take more structure — usually founder-led or owner-led, with more potential than momentum, and with real operational complexity: multiple locations, a technical product, a long sales cycle, or a catalogue nobody outside the building fully understands. Six industries where the pattern shows up most: manufacturing and automation, distribution, food service and restaurants, consumer goods, commercial services and trades, and M&A firms with their portfolio companies. Whether a particular business is a fit is a short conversation, not a revenue threshold.

The profile that fits.

Privately held. Founder-led, owner-led, family-held, or PE-backed. The common factor is that decisions can be made by people who are actually in the conversation.

More potential than momentum. Something real has been built — a product, a reputation, a customer base — and growth has flattened or become harder than it should be. The business has outgrown the way it sells, communicates, and delivers.

Operationally complex. This is the qualifier that matters most and it is not about size. Multiple locations. A technical product requiring explanation. A long, multi-stakeholder sales cycle. A catalogue with thousands of SKUs. Regulatory or engineering constraints. Complexity is where the promise gap hides, because complexity is what the founder used to hold together personally.

Six industries: manufacturing and automation, distribution, food service and restaurants, consumer goods and D2C, commercial services and trades, and M&A firms with their portfolio companies. The process does not change between them — we just already speak the language, know the buying cycle, and know where the gap usually opens.

What actually predicts a good engagement, more than industry or size: a leadership team willing to be told the problem is not marketing. That is the real qualifier, and it is why fit is a conversation rather than a threshold.