Strategic Context for Agencies and Consultants

A positioning consultant we work with kept thirty-one client engagements in her head, a Notion workspace, two Google Drives, and a folder on her desktop called __archive_REAL. When a former client called eighteen months after their engagement asking for a refresh, she billed four hours of re-onboarding before she could answer the actual question. The work she'd done — the interviews, the competitive map, the rejected category nouns — was somewhere. Finding it cost more than redoing it.

This is the agency tax on strategy work. Each engagement is treated as a clean slate, even when it isn't. Each new analyst on the account starts from the deck, not the reasoning. And the third time a client asks "didn't we decide something about this?" the honest answer is: we did, and I can't find it.

3.1x

The compounding problem

Most agencies and independent consultants operate on what we'll call the artifact model: every engagement produces a deliverable — a positioning deck, a messaging house, a brief — and the deliverable is the institutional memory. Six months later, the deliverable is the only thing left. The reasoning behind it, the alternatives that were rejected, the customer voice that informed it, the competitor moves that shaped it — all gone, or scattered across Slack threads and Loom recordings nobody indexed.

The artifact is a snapshot. The snapshot doesn't compound.

What compounds is context — the structured record of what's true about a client's market, where their positioning sits relative to alternatives, what's been tried, and what changed. When context is captured and reusable, every engagement after the first costs less to start and produces sharper work.

The deliverable is the snapshot. The context is the asset.

What "strategic context" actually means

Strategic context, as a working artifact, has four layers. Each layer answers a different question for the practitioner who picks up the account next quarter — even if that practitioner is you, eight months from now, with no memory of the original conversation.

The artifact model captures Layer 1 (sometimes) and ignores 2, 3, and 4. The context model captures all four, in a structure that survives staff changes and tool migrations.

Why agencies underbuild this

Three reasons keep coming up in our conversations with agency principals and independent consultants.

The economic incentive points the wrong way. Building reusable context takes time inside an engagement that the client isn't paying you to build a system. Most agencies treat it as overhead and never invest. The cost shows up later, on the next engagement, but it's diffuse enough that nobody traces it back.

Tools are biased toward delivery, not memory. Figma is for the deck. Notion is for the project plan. Slack is for the conversation. None of them are structured for "what's true about this client's market in April 2026, and how is that different from what was true in October 2025?"

Client confidentiality makes the easy answer worse. You can't dump everything into one shared workspace because one client shouldn't see another. So agencies create a folder per client, and each folder reinvents the structure from scratch.

MR

We did seventeen engagements last year. I can tell you what we shipped for each one. I cannot tell you, without digging for an hour, why we rejected the original category noun for the fintech client in March. That reasoning is somewhere in a Loom transcript.

Build the context system once, then run it

The fix is structural. Build a multi-client context template — one shape that every engagement fills out — and a discipline for keeping it current. The template is the asset. Every individual client file is an instance.

The template costs about two days to build and saves roughly eight hours per returning engagement after the first. For a four-person consultancy doing fifteen engagements a year with a 40% repeat rate, that's roughly 48 hours back annually — a week of senior time you weren't getting paid for.

What the template needs to capture

Below is the minimum viable structure. If your engagements are heavier on launch work or pricing, you'll add layers. The four below are the floor, not the ceiling.

Per-client context file: minimum sections

The economics, briefly

We pressure-tested this with three independent positioning consultants and two small agencies. The pattern was consistent: clients with structured context files were re-engageable at roughly 60–70% of the original engagement cost, because the agency wasn't rebuilding the foundation. Clients with no structured context were re-engageable at 90–100% of original cost, because the agency genuinely had to redo the discovery work.

The clients didn't know which bucket they were in. They just experienced one agency as faster and sharper on the second engagement, and the other as starting from scratch.

The third time we re-engaged with the same client and I had to ask the same questions I'd asked eighteen months ago, I realized we weren't running an agency. We were running a series of forgetful relationships.

Where this lives in your stack

You don't need a new tool to do this. Most agencies start with a structured Notion or Coda template, one page per client, locked sections matching the four layers. The constraint is the discipline, not the surface.

Two failure modes to avoid. First, don't build the template inside the same workspace where you do delivery — the client deck and the client context need different lifecycles. The deck ships and freezes. The context updates. Second, don't let any single client file drift from the canonical template structure. The moment one engagement's file looks different from the others, the system is dead. The template is the contract.

For agencies that have crossed roughly twenty active clients, the manual approach starts to break — quarterly reviews slip, signal feeds go stale, customer-voice libraries become unreadable. That's the threshold where a structured context system with monitoring (a feed that surfaces market changes per client without you having to remember to look) starts paying back.

What to do this week

Pick your three highest-revenue clients. For each, spend forty-five minutes filling in the four-layer template from memory and existing artifacts. You'll discover which layers are robust and which are wishful thinking — usually the decision log is the one with the largest gap. Schedule a thirty-minute interview with the account lead to fill it in while it's still recoverable.

Then commit to a rule for the next engagement you start: the template gets filled in during the engagement, not after. The first time you do this, it adds maybe two hours to the project. The third time a client comes back, you'll get those two hours back in a single afternoon.

The agencies that compound aren't the ones with better frameworks. They're the ones that remember what they decided, why they decided it, and what was true about the market on the day they decided it.