# Stop Sending Reports. Build a Forecast Your Client Can Argue With.

Most agencies report backwards, framing every number after the fact. A forecast your client helped build turns the monthly call from defence into diagnosis.

Published: 2026-07-25

Source: https://grey-wolf.co.za/insights/build-a-forecast-your-client-can-argue-with/

---

This month I sat on a call watching a client's ops lead argue with my numbers.

Best thing that happened to that account all month.

We'd built them a lead-forecasting tool. Not a spreadsheet lobbed over the wall in an email. A working model, on screen, live: scenarios across every product line, cost-per-lead targets per product, seasonality built from five years of their own search data. The August scenario we landed on mapped about R800k of spend to just under 6,000 leads at a blended R139 a lead.

Their ops lead didn't buy one product's seasonality curve. Their numbers person poked at a conversion assumption. Fine. We adjusted, re-ran the scenario while they watched, and the forecast that came out the other side was partly theirs.

Theirs. Sit with that word, it's the whole trick.

Two days later, the forecast got tested.

## The week the wheels wobbled

Friday's report: blended cost per lead running 31% over forecast. Two products on zero leads.

On paper, a shocker of a week for the agency. Time to polish up the excuses, right?

Except here's what actually happened. Because the forecast existed, the drift was visible within days, not at month-end. Because the client helped build the baseline, nobody argued about the baseline. The call went straight to causes. One product had a landing page change mid-flight. Another had a tracking gap we'd already root-caused. We spent the whole call on fixes.

Not one minute defending our existence. Not one.

Now compare that to the standard agency month-end. A report arrives. The numbers are whatever they are. The client reverse-engineers whether to be upset, and the agency reverse-engineers a story for every metric, because nothing was ever promised in advance.

That's not reporting. That's defence. And clients smell it a mile off.

## Why agencies won't commit to a number

Fear. Plain and simple.

A forecast is a promise, and a promise can be missed. So the industry defaults to reporting backwards, where every number can be framed after the fact and nobody's ever technically wrong.

I think that's exactly backwards. Hope is not a strategy, and neither is hindsight.

The agencies that get churned are the ones whose value nobody can pin down. If your client can't say what good looks like, sooner or later someone cheaper walks in with a story that sounds just as plausible as yours. A forecast makes your value testable. Uncomfortable? Very. It's also the only reason hitting it means anything.

And missing it means something useful too. A 31% miss against a shared forecast is a diagnosis. A 31% anything against no forecast is just a Friday.

## Want to copy this?

Four rules.

Build it from the client's own history. Five years of their search and conversion data beats any industry benchmark PDF you'll ever download.

Forecast per product, then blend. Blended-only numbers hide exactly the failures you need to catch fast. Our two zero-lead products were completely invisible in the blend.

Build it with them, live. A forecast you deliver is your forecast. A forecast they argued over is theirs, and they'll defend it in their own boardroom when you're not in the room.

Then report against it weekly, and every report becomes one question: are we where we said we'd be, and if not, why?

Clicks are what happened.

A forecast is what you're for.