Strategic Context for Agencies and Consultants

The third time you write a positioning brief for a vertical SaaS company in healthcare, you are doing something embarrassing: you are starting from a blank page. The patterns you found in client one and client two — the ones that took weeks of interviews, deck audits, and pricing-page archaeology to surface — are sitting in a Google Drive folder you will not open. Your next associate will rediscover them. Your next proposal will quote the same hours.

This is the agency tax on strategy work. And it is almost entirely self-inflicted.

By "strategic context" we mean the working layer of an engagement — the win/loss patterns, the competitor reads, the message hypotheses, the buyer-language inventory — that an analyst builds up over the first three or four weeks and then, traditionally, throws away when the SOW closes.

62%

Why client work doesn't compound (and the standard fixes don't fix it)

The usual answer is "build a knowledge base." Agencies have tried this for twenty years. Notion wikis. Confluence pages. Internal newsletters. They fail for the same three reasons.

Reason one: the unit of capture is wrong. A wiki page about "what we learned at HealthCo" is too coarse. The next strategist needs the question that produced the insight — not the slide that summarized it. Without the question, the insight is decorative.

Reason two: confidentiality kills capture. A specific finding ("HealthCo's win rate against Epic is 12% in mid-market") cannot be written down anywhere a future client could see. So nothing gets written down. The pattern dies with the engagement.

Reason three: the senior strategist is the knowledge base. Pattern recognition lives in one head. When she is on a deck deadline — which is always — the junior strategist starts over.

A working context layer has to solve all three at once. It captures questions and abstractions, not findings. It separates the pattern from the evidence. And it gets updated by the people doing the work, not by a librarian.

The deliverable is the byproduct. The pattern library is the asset.

The compounding context model

Here is the shape of the system. Three layers, kept deliberately separate.

Layer 1 you already have. It is the SOW deliverable.

Layer 2 is what you should be building every week. One paragraph per pattern, no client names. "In healthcare SaaS with an ACV above 80K, the security review stalls more deals than pricing — and almost no founder believes this until the third lost deal." That is a portable observation. It is also a sentence you can drop into a future kickoff to demonstrate fluency.

Layer 3 is the multiplier. The reason a senior strategist outperforms a junior one is not that she knows more — it is that she asks better questions in week one. Capture the questions.

Building the system without a six-month rollout

Most agencies that try this announce a knowledge initiative, hire a content lead, build a taxonomy, and quit by month four. Don't do that. Build it in the margins of work you are already doing.

What confidentiality actually permits

The reflexive answer is "we can't share anything." The accurate answer is more useful.

You cannot share: client names, financial data, specific customer lists, named individuals, proprietary product roadmaps, anything covered by an explicit NDA clause.

You can share: the shape of the problem, the pattern of the solution, the diagnostic questions that worked, the category-level observations, the workflow you used. You also can — and should — clear unusual cases with the client. Most are flattered to know their engagement informed a portable framework. Few read the line in your MSA that already grants you this.

DO

We had eighteen years of work in folders nobody opened. The thing that broke it loose was deciding that the unit of capture was the question, not the answer. Questions don't violate NDA. Questions are also what the new hires actually need.

What to capture, by engagement type

Different engagement types yield different patterns. A positioning project will produce category and message patterns. A demand-gen audit will produce funnel and channel patterns. A win/loss study produces buyer-language and objection patterns.

Capture targets, by engagement type

The pattern library should not try to cover everything. Pick the two or three engagement types you do most often. Saturate those first. Add a third in year two.

How this shows up in the proposal

Here is the commercial point. A pattern library is not an internal nicety. It is a sales asset.

When a healthcare-SaaS founder asks why your boutique agency should win the work over a McKinsey alum or a freelance specialist, the honest differentiator is depth in the shape of his problem. "We've done eleven engagements with vertical SaaS companies between 5M and 30M ARR. Here are the three patterns we expect to see in your data, and here are the questions we'll ask in week one to confirm or kill them." That is a credible opening. A blank deck is not.

The library also changes how you scope. The fourth time you do an engagement of a given shape, you should be 30–40% faster and producing better work. Some of that speed should reach the client as a lower price; some should accrue to your margin. Both are correct. Neither is possible if every engagement starts from zero.

We charge more than we did three years ago and the engagements take less time. The bridge is the question bank. We walk in knowing what to ask.

The template, and how to use it

The Multi-Client Context Template below is a structured Word doc with three sections matching the three layers. Use it once for a recent engagement to get the feel. Then add it to your standard project-close checklist — forty-five minutes of structured retro before the team rolls off.

What to do this Friday

Pick the engagement that closed most recently. Spend forty-five minutes with the lead strategist. Answer the three retro questions, abstract the answers into three or four portable sentences, and write the diagnostic questions that produced them. File it in a single document — anywhere, the location matters less than the existence.

Do this for the next three engagements that close. By the fourth, you will start recognizing patterns across them. By the eighth, your week-one kickoffs will look different. By the twentieth, the library is the thing your associates open before they open the brief.

The strategy work was always compounding. The capture was the missing piece.