Day 132: Your Case Studies Can Teach AI the Wrong Customer Profile
A company can publish accurate case studies and still present a misleading picture of whom it serves.
The problem is not truth inside each story. It is the shape of the visible sample.
Perhaps the company built its reputation through large enterprise transformation programmes. Those projects generated recognisable logos, approved testimonials, and polished narratives. They became the public portfolio.
The business has since developed a lighter delivery model for mid-market teams. That offer is current, commercially important, and better suited to the buyers leadership wants next. Yet the public examples still show global rollouts, multi-quarter change programmes, complex procurement, and large implementation teams.
A suitable prospect may reasonably conclude that the company is too expensive, too complex, too enterprise-focused, or simply not designed for their situation. An answer-led system encountering the same accessible examples under recorded conditions may also frame the company around the segment that dominates the visible portfolio.
That does not mean the system has discovered the company's ideal customer profile. It means the public sample made one interpretation easier than others.
For CMOs, Marketing Directors, and founders, the distinction is commercially important: the customers you have served, the examples you can publish, and the buyers you want next are three different sets.
The public portfolio is a selected sample
Case studies are never a complete census of customer work.
They pass through selection pressures. Some customers grant publicity rights and others do not. Some outcomes are easier to explain. Recognisable brands are more likely to receive design and editorial budget. Large engagements create more material than small ones. Older projects may have polished assets while newer offers are still waiting for a publishable example.
The visible portfolio therefore reflects permission, incentives, history, and publishing capacity as much as the actual customer base. That is normal commercial publishing.
The risk appears when leadership treats this selected sample as a neutral representation of the market served now. Buyers and public systems cannot see the unpublished work or hear the private explanation that the delivery model has changed.
If one segment dominates the accessible material, repeated context can become the easiest summary.
A generalised portfolio-skew teardown
Consider a fictional B2B operations consultancy. This is a generalised illustration, not an observed client or platform result.
The consultancy has served three broad customer profiles:
- global enterprises commissioning multi-region transformation programmes;
- established mid-market companies buying a bounded diagnostic and implementation sprint;
- specialist teams needing a focused intervention before handing delivery back in-house.
Its public portfolio contains six detailed stories. Five feature global enterprises. Four mention steering committees, multiple business units, and long implementation programmes. The sixth is a partner announcement rather than a delivery example.
The published material is accurate. The portfolio is still skewed.
Now imagine a Marketing Director at a 250-person company evaluating the consultancy. She needs a four-week diagnostic, has a small internal team, and wants to avoid a transformation programme. The public examples may lead her to ask:
- Is this company interested in an engagement of our size?
- Will the work require enterprise procurement and a large internal programme team?
- Is a diagnostic available on its own, or only as the opening phase of a wider transformation?
- Are the showcased results achievable only for companies with global data and budgets?
None of those conclusions has been stated. The portfolio has supplied enough repeated context for them to be plausible.
An answer-led comparison could face the same ambiguity. Accessible examples may support a description such as “an enterprise transformation consultancy for complex, multi-region programmes”. That may fit the visible cases and still be incomplete for the current offer. Different systems and conditions may produce different summaries; this illustration does not claim universal behaviour.
The commercial issue is the gap between three maps:
| Map | What it contains | What it does not prove |
|---|---|---|
| Customers served | The real history of organisations and engagements the company has delivered. | Which examples can be disclosed or which segment the company wants next. |
| Customers shown | The selected examples visible across the public web. | The complete customer mix, minimum engagement, or current delivery model. |
| Customers wanted next | The buyer profiles leadership has chosen to pursue now. | That the company already has permission to publish a matching example. |
A portfolio problem exists when those maps diverge without explanation.
Accuracy is necessary, but representation needs context
The wrong response is to manufacture a “representative” case study for a target segment the company has not served.
Do not invent a customer, inflate a small engagement, imply an outcome that was never measured, or recast an enterprise project as mid-market work. A desired customer profile is a commercial direction, not retrospective evidence.
The responsible options are narrower.
First, improve context around existing stories. If an enterprise case involved a bounded pilot before the larger programme, say so only when the approved record supports it. If the method can be purchased at a smaller scope today, describe that current offer outside the historical case rather than rewriting the past.
Second, publish truthful forms of relevance that do not pretend to be customer proof. A fit page can state the intended buyer, prerequisites, typical scope, and non-fit conditions. A delivery page can explain the current route without attaching it to an invented outcome. An anonymised example may be possible where permission and verification allow, but it must not become composite fiction.
Third, treat the gap as a commercial decision. If the company wants a segment it has not yet served credibly, the answer may be a pilot, a revised offer, or deliberate business development rather than more assertive copy.
The goal is not to make every segment look equally represented. It is to stop the public sample silently standing in for the whole business.
Run a portfolio-skew audit
A compact audit can reveal whether the visible mix is helping the right buyers recognise themselves.
List every accessible customer story, logo page, partner profile, award entry, and substantial third-party description that a buyer may encounter. For each item, record only what the published material can support:
| Dimension | Audit question |
|---|---|
| Company size | Which size or complexity cues are explicit, and which are merely implied? |
| Buying maturity | Does the example assume a first diagnostic, an established programme, or a mature operating model? |
| Engagement burden | What team, timeline, integration, or governance load appears necessary? |
| Sector and geography | Does one industry or market dominate the visible set? |
| Procurement route | Do the examples imply enterprise tendering, partner-led procurement, self-serve purchase, or a direct advisory engagement? |
| Offer and outcome | Which current offer does the story genuinely support, and where does the evidence stop? |
Then compare the distribution with the buyer profiles leadership wants next.
Do not turn the exercise into a quota. Ask a commercial question instead:
What wrong assumption could a suitable buyer make because the visible examples overrepresent one part of our history?
That assumption might concern minimum engagement size, implementation effort, sector fit, geographic reach, buying maturity, or the need for enterprise procurement. Choose the one with the clearest consequence for current positioning.
The resulting action may be to add context to an existing story, clarify present-day offer fit, commission a truthful new case when suitable work and permission exist, or accept that the target segment is not yet supported strongly enough to lead with.
Keep the GEO claim bounded
Clear segment, scope, and delivery context may reduce avoidable inference for buyers and for answer-led or search-led systems encountering the public material under recorded conditions. It cannot guarantee inclusion, ranking, citation, recommendation, buyer behaviour, attribution, conversion, or revenue.
A portfolio audit also cannot prove why a particular answer used a particular customer profile. The responsible observation is limited to what material was accessible, what pattern appeared in recorded checks, and which interpretation the company made easy or difficult.
For Google AI features, ordinary Search fundamentals remain the foundation. Google does not require llms.txt, special AI markup, arbitrary chunking, or over-focused structured data as switches for AI visibility. Accurate, useful, accessible pages for people and core Search remain the priority.
Decide whether the portfolio matches the next market
If every public example points to the same company size, implementation burden, sector, geography, and procurement route, the portfolio tells a coherent story. Leadership should confirm that it is the story the business wants to tell now.
If it is not, do not rewrite history and do not manufacture balance.
Separate the customers served from the customers shown. Compare both with the buyers wanted next. Identify the most consequential skew, add accurate context where the record allows, and make an explicit choice about the missing representation.
Every case study can be true while the portfolio still points at the wrong market.
The next publishing decision should correct the sample. Otherwise, acknowledge what the business must earn before it can.