The symptoms were familiar. Inbound had plateaued. Two SDRs were sending a respectable volume and booking almost nothing. Paid was running, spending steadily, and reporting a cost-per-lead that leadership had stopped believing. Nobody could say which of the three was working, because nothing was measured the same way.
What follows is the 90-day sequence we ran. The numbers are directional and the client is anonymised, but the shape is typical.
Days 1–14: audit, and stop the bleeding
The first two weeks are not building. They are finding out what is true.
- Sending domains were healthy but all traffic ran through the primary corporate domain — one bad campaign from a serious problem.
- The "ICP" existed as a slide. The actual list being sequenced matched it on roughly a third of records.
- Paid audiences were uploaded quarterly by hand and had drifted a full quarter out of date.
- Roughly one in five accounts in outbound had an open opportunity or was already a customer.
The last one is the common gut-punch. Fixing suppression alone recovered credibility with the sales team within a fortnight.
Days 15–35: rebuild the audience as a query
We replaced the static list with a live definition — headcount band, two tech signals, hiring activity in specific functions, funding recency — and ran it against a properly enriched index. The qualifying set came out about forty percent smaller than the old list and materially better.
That is the trade almost every team resists and almost every team should take. A smaller, correct audience beats a larger, hopeful one on every metric that matters, including cost.
A smaller, correct audience beats a larger, hopeful one on every metric that matters.
Days 25–50: put paid and outbound on the same audience
This was the change with the largest single effect. The same account list now drove both the sequences and the ad audiences, synced daily rather than quarterly.
The mechanism is not complicated. An account entering outbound also starts seeing the brand. By touch three, the sender is not a stranger. Reply rates on the sequences rose noticeably, and the ad spend stopped being evaluated on its own — it became a cost of making outbound work, which is a much easier conversation with a CFO.
Days 40–70: fix the handoff
Interested replies were taking an average of eleven hours to get a human response, and a meaningful share were being classified wrongly by a human skimming an inbox. We put classification and routing on rails, offered times in-thread, and wrote outcomes back to the CRM automatically.
Median time to first response went to under twenty minutes. Nothing about the messaging changed; the booking rate moved anyway.
Days 60–90: one attribution model
Last step, and the one that made the programme defensible internally: a single view where an ad impression, an email reply and a booked call all attach to the same account, and cost-per-opportunity is computed blended rather than per-channel.
It ended the quarterly argument about whether paid or outbound "worked". The answer, predictably, was that neither worked alone and together they worked well.
Where it landed
By day 90 the programme was producing meetings on a weekly cadence that leadership could forecast against, at a blended cost-per-opportunity roughly a third below where the two channels had been running separately. The SDR team stopped building lists and spent their time on live conversations.
The part worth stealing is not any single tactic. It is the order: fix what is true, then the audience, then the channels, then the handoff, then the measurement. Doing it in that order is why it took a quarter instead of a year.