The ad everybody loved
Tessellate ran five creatives against a 210-account list of North American data teams. One of them — a single-image ad with a strong stat in the headline and a link to an ungated benchmark report — was carrying the programme on every metric anyone was looking at. Roughly 44% of clicks, the lowest cost per click by a wide margin, and a click-through rate about twice the next best ad.
It was also the ad the team was proudest of, which matters more than it should.
What cost per account moved showed
Kairo scores accounts on committee breadth, depth, seniority and recency, and records which creatives an account was exposed to before each stage change. Attributing stage movements to the creatives that preceded them produced a table nobody was expecting.
| Creative | Share of clicks | Accounts moved | Cost / account moved |
|---|---|---|---|
| SI-01 — benchmark report | 44% | 6 | $1,840 |
| SI-04 — migration cost | 12% | 19 | $290 |
| DOC-02 — architecture guide | 9% | 14 | $355 |
| SI-02 — customer quote | 21% | 4 | $2,110 |
| VID-01 — product tour | 14% | 7 | $720 |
The benchmark report was extremely good at getting clicks from data engineers at companies that were not on the target list and were never going to buy. It was doing what a well-written ungated resource does: attracting an audience. It was not doing what an ABM creative has to do, which is get several people at one named account to take the problem seriously.
“Cost per click made one ad look like our best. Cost per account moved made it look like an expensive way to entertain people who will never buy.”
What they changed
- SI-01 and SI-02 were paused. Together they were taking 65% of clicks and producing 22% of account movement.
- SI-04 and DOC-02 took the budget, but at a lower total: the programme went from roughly $14,000 a quarter to just under $10,000, because the two efficient creatives saturated the reachable list before the old budget was spent.
- A frequency floor replaced a spend target. The goal became “every Tier 1 and Tier 2 account sees SI-04 at least four times in six weeks” rather than “spend $14k”.
- The benchmark report moved to organic and email. It was genuinely good content; it was just being paid for as if it were an ABM asset.
The measurement change that made it possible
None of this required new data. LinkedIn already knew which creative each impression carried. What was missing was the join between creative exposure and account stage movement — and, more importantly, a metric the team had agreed to steer on before the numbers arrived.
Tessellate now reviews three numbers weekly: accounts moved by stage, cost per account moved by creative, and Tier 1 coverage. Clicks are still recorded. They are no longer discussed.
What they would do differently
- Test creatives against account movement from the start. Two quarters of creative iteration had been optimising against the wrong outcome, and the iteration was good — it just made SI-01 better at the thing that did not matter.
- Stop treating budget as a target. Once you know the cost of reaching your list at the frequency you want, spending more than that is a decision to reach people who are not on it.
- Keep one high-CTR ad running anyway. They kept VID-01 for the top of the funnel. Not everything has to be measured on account movement; it just has to be honest about which job it is doing.