What is the Recall Tax in consulting?
The Recall Tax is the cost of recovering context your firm already has. Every hour your people spend reconstructing something the firm already knows — what was said, what was promised, where the document went — is Recall Tax. The defining word is already. It isn't the cost of not knowing. It's the cost of knowing and not being able to reach it in the thirty minutes that matter.
That distinction rules out the wrong fix. The Recall Tax disguises itself as a discipline problem — tag better, name files consistently, be more organized — and that instinct sends firms into months of process-hygiene that doesn't touch the cause. The tax isn't produced by undisciplined people. It's produced by the architecture of how a firm's tools store information, and disciplined people inside that architecture pay it just the same.
How do consulting firms lose client context they already have?
They don't lose it — they store it across tools that were each built for a different job, with no shared spine connecting them. Every tool in the stack is a silo of memory, and in combination they guarantee the tax:
- Slack is a river. Organized by recency; everything flows downstream and out of sight. Search returns forty-two keyword hits, not the one answer to your actual question.
- Email is a set of private drawers — one per person, partitioned by who happened to be on the thread. The largest store of what a firm told its clients, and the most likely to walk out the door.
- The drive is a filing cabinet whose taxonomy is a fossil. The document is there; finding it means knowing the mind of whoever built the folder structure — who has usually left.
- The task tool knows the what and never the why — and the calls are gone unless someone transcribed them, and the transcripts are gone unless someone read them.
Each tool is well-built for its own job. None is broken. And yet the firm's memory — the union of all of them — is unreachable as a whole. The firm has five memories and no memory.
How much does the Recall Tax cost?
For most firms that count honestly, recall lands between a fifth and two-fifths (20–40%) of senior time — roughly four to ten hours per senior, per week. That range isn't borrowed from a study; it's what falls out of the arithmetic when firms actually count.
A worked example
≈ $78,000/year
5 seniors × 3 hours/week each spent recovering context the firm already had × $100/hr blended cost × 52 weeks. Three hours sits below the 4–10 hr range most firms find — so for a ten-person firm this is the conservative number, not the alarming one.
The point isn't the exact figure — it's that the honest number is always large enough to be uncomfortable, and it grows every time the firm adds a tool, a person, or loses a tenured one.
Why can't you see the Recall Tax?
Because it has no line item and lives inside the firm's healthiest-looking number. When a partner spends thirty minutes reconstructing an account, the firm's instruments record — if the time was billable — thirty minutes of billable senior work on a client, exactly the number the firm wants more of. The single largest diversion of the firm's scarcest resource is the one thing its dashboard is built not to show. The only way to get the number is to measure it directly, once, by hand.
Why does the Recall Tax get worse as the firm grows?
It grows along the three axes that also describe a firm getting more successful:
- Tools. Every new tool adopted to solve a problem is a new silo — another slice of memory behind another set of terms. Adding tools increases the tax.
- People. A two-person firm has almost no Recall Tax — both people were there for everything. At twenty-five, no one has witnessed more than a fraction, so recall becomes a constant cost of operating as one firm.
- Turnover.Every senior who leaves takes their personal index — their map of where things are and what was decided. Their departure converts a slab of the firm's reachable memory into tax, in a single afternoon, with no event on any dashboard.
Can you just throw AI at the Recall Tax?
Not generic AI. A capable model pointed at the firm's data that produces an answer you still can't trust just lets you be confidently wrong faster: you either re-verify by hand (paying the tax anyway) or walk into the renewal call carrying a machine's unexamined guess. The tax is only retired by an AI grounded in the firm's real sources and able to show where every answer came from — which turns a thirty-minute reconstruction into a five-second, verified retrieval. That difference — citation, not just summary — is the whole game.
How do you measure the Recall Tax on your own firm?
You can do it in a week, and crude-and-real beats precise-and-theoretical. For one week, have each senior keep a single running tally — one line on their phone each time they catch themselves recovering something the firm already knew:
brief before Acme call — 30m · hunting for the SOW version — 15m · reconstructing what we told Beta in Q1 — 25m
At week's end, total the minutes per person, convert to hours, multiply by their hourly value, sum across seniors, and annualize. That's your Recall Tax — measured, not guessed, and every line in it a real instance someone logged. It will be larger than your estimate and harder to argue with, because you can't dismiss your own week.