Last week I wrote about the gap between the version of your business that exists in the data and the version actually being built. Several of you replied with the same follow-up question, in different words.

‍

"Fine. But how do I take that into a budget conversation?"

‍

Good timing. It's late July, which means autumn spend plans are landing on board agendas everywhere. So this week: what separates a budget the board respects from one they pick apart line by line.

The wish list problem

Most marketing budgets are activity lists with prices attached. Paid search: £X. Events: £Y. Content: £Z.

‍

It looks sensible. Then one number misses, and suddenly the whole plan is up for renegotiation.

Because nobody agreed in advance what each number was supposed to prove.

‍

A budget the board respects is a model. The assumptions are on show, the numbers have jobs, and everyone knows in advance what would change the plan.

What I'm building this week

I'm putting together Q4 spend projections for a client right now, so this is live for me too.

‍

Here's what goes into it.

‍

CAC by channel, blended only for an overview of full activity.

‍

And a ceiling for every channel. Each line in the plan carries an estimate of how much spend that channel can absorb before efficiency decays. Doubling budget on a channel near its ceiling buys you a rounding error.

‍

The board should see that maths before they ask for it.

‍

Trigger points are for every material line: if this hasn't happened by this date, the money moves there. Reallocation rules agreed before the spend starts, when everyone is calm, rather than mid-quarter when nobody is.

‍

And ensure there’s space to test. Don’t let your budget only allow for the known quantities. You’ll miss the magic.

The £500 that earned its keep

Last week the same client spent £500 on competitor keyword campaigns. Eight ad groups. Zero conversions.

‍

In a wish list budget, that's a failure to bury before the next review. In a modelled budget, it's a cheap answer. We now know that route doesn't work at this stage, we found out for £500 rather than £5,000, and the money has already moved.

‍

Boards don't punish tests that fail. They punish spend nobody can explain.

‍

If your budget has no lines that are allowed to fail, it has no learning in it. Which means every pound is a promise, and promises get renegotiated the moment reality wobbles.

What a board actually respects

Nobody who has read a forecast before believes one to the pound. What they're checking for is whether there's a thinking system behind the numbers.

‍

Can you explain why the money is weighted the way it is? Do you know where your ceilings are? Have you decided, in advance, what evidence would make you move budget, and when?

‍

A founder who can answer those three questions with a smaller budget will beat a bigger budget defended with confidence alone. Every time.

The one thing to do this week

Take your three biggest line items and write one sentence next to each: the result that would make you cut it, and the date you'd decide by.

‍

If you can write those sentences, you have a budget you can stand by. If you can't, you have a hope with a number next to it.

‍

Drop me a line if you're building autumn plans right now. I'm deep in exactly this at the moment and happy to compare notes on where the ceilings tend to hide.

‍

PS. This is also the work our portfolio CMOs quietly do before a founder's board meeting: turning an activity list into a model the founder can defend without flinching. If your next board pack has a budget slide you're not looking forward to, reply. It's usually a faster fix than people expect.