What a 70/30 split costs in 18 months of CPC inflation

The 70/30 activationConverting people who already want the thing, through search, retargeting and promotions. It captures demand rather than creating it.-to-brand split is the number almost every account lands on without anyone deciding it. Quarter by quarter, budget drifts from the brand line to the measurable ones. Binet and Field1 put the healthier number nearer 60/40 the other way, brand ahead of activation, but that is a fight for another essay. This one is narrower. If you run an activation-heavy split for eighteen months, what does it cost you, measured in the simplest number on the account, the cost per click?

The mechanism, in one paragraph#

Cut brand spend and your share of searchYour brand's share of all searches in your category: how often people look for you against everyone you compete with. falls. Branded clicks are cheap; generic clicks cost several times as much. As the mix shifts from branded to generic, your blended cost per clickWhat you pay each time someone clicks an ad. climbs, and it compounds.

I laid the mechanism out in full in an earlier essay, so here is the short version. Brand spend keeps demand warm, and warm demand shows up as people searching your name. Those branded clicks are the cheapest and highest-converting traffic you own. Cut brand spend and, over a few quarters, your share of category search slips. The searches you have lost were the cheap ones, so you replace that volume on generic, non-brand terms, where you bid against everyone and every click is someone who has never heard of you. Your blended cost per click rises. Nothing changed in the auction. The mix moved from cheap clicks to expensive ones.

The eighteen-month arithmetic#

Put rough numbers on it. This is a model, not a measurement. Say branded search is 40% of your clicks today at £0.30 each, and generic is 60% at £1.80. Your blended cost sits at £1.20. Now run this split for six quarters. You lose a few points of share of search a quarter, so branded volume falls from 40% of the mix toward 28%. Generic makes up the gap. Nothing dramatic happens in any single quarter. But the cheap share is shrinking, and the blended cost rises with it. By month eighteen the same clicks cost £1.38, 15% more, for demand you used to get for less.

The exact curve is yours. Pull your branded and generic click shares and their average costs, decide how fast your share of search is moving, and run the same blend forward in the model at the top of this page. The number that comes back is a floor on the cost of the split: it only counts the mix effect and ignores the conversion-rate drag that a colder audience adds on top.

The branded share is the cheap share. Lose it and the blend costs more every quarter.

What the number leaves out#

The model only counts the mix effect. It ignores the conversion-rate drop from colder traffic, and rebuilding share of searchYour brand's slice of all the searches in your category: how often people look for you against everyone you compete with. takes longer than losing it. The real cost is higher than the curve.

The true cost is worse than the chart. The colder generic traffic converts worse than the branded traffic it replaced, so the same spend buys fewer customers before the cost per click even moves. And the timing is not symmetrical: losing share of search takes quarters, and winning it back means rebuilding demand from a lower base, which takes longer than the fall. The share paths behind the buttons are the assumption to test against your own share-of-search history. They also key everything to the split alone, when what matters in practice is your weight against the category, so a market leader and a challenger on the same split will not share a fate. The eighteen-month figure is only the start of the cost; the rest sits with the triangle’s other two corners, which this model does not price.

Eighteen months apart#

In the model nothing moves for the first two quarters, and a live account is noisier still, which is why the split survives review after review. The cost lands eighteen months after the decision, in a different quarter, in a different planning cycle, and it lands on performance.

You cut brand to protect performance, and eighteen months later performance is the thing paying for the cut.

So run your own eighteen months before the next planning round. If your cost per click bends up while the category’s does not, the split is the first suspect, and more budget at the same split only buys the bend more slowly.