Any single buying signal has a weak relationship with actually buying.
Someone engaged with a competitor’s post. Fine, they might be evaluating, or they might like the author, or they might have tapped a reaction while scrolling. A company posted a job. It might mean growth, or it might be a backfill nobody thought about twice.
Act on single signals and you get a list with a lot of noise in it. Act on combinations and the hit rate changes substantially, because two independent indicators pointing the same way is a much less likely coincidence.
Why stacking works
Each signal has a false positive rate. Post engagement is high, most people engaging with a post about your category are not buyers. Job postings are moderate. Job changes at ICP-fit companies are lower.
When two signals are independent, driven by different underlying causes, their false positives do not overlap much. The chance that both fired coincidentally is considerably lower than either alone.
The word doing the work is independent. Two signals measuring the same thing stack poorly. Someone engaging with three competitor posts in a week is one signal observed three times, not three signals.
Combinations that work
Hiring surge + leadership change. A company hiring three people into a function that just got a new head. The new leader is building a team and will buy tooling to support it. Among the strongest combinations available.
Job change + ICP-fit new company. The move alone is a signal; the move to a company that matches your ICP is the real filter. Most teams treat these as one thing when they are two.
Competitor engagement + hiring in the relevant function. They are looking at alternatives and expanding the team that would use it. Evaluation plus capacity.
Topic post + seniority. A director writing about a problem is a different proposition from an analyst doing the same. The post is the signal, the seniority is the multiplier.
Promotion + past evaluation. Someone who looked at you before and now has authority. This is the highest-converting stack in B2B and it uses no external data, both halves are in your own CRM.
New search match + headcount growth. A person newly matching your criteria at a company that is also growing. Weak on its own, decent together.
Combinations that do not
Multiple engagements from the same person. One signal, repeated.
Hiring + headcount growth. Both measure the same underlying thing. Stacking them inflates the score without adding information.
Any signal + a company size that matches your ICP. ICP fit is not a signal, it is a filter. Treating it as a stacking input double-counts it, because you should already be filtering on it before anything reaches a score.
That last one is a common and quiet error. If your scoring model gives points for being in-ICP and filters for in-ICP, you have built a model where everything scores well.
Scoring without overbuilding
The temptation is a weighted model with a dozen inputs. Resist it, you will not have the volume to validate it, and an unvalidated weighting is just a guess wearing a number.
A workable approach:
Tier the signals.
- Strong, job change to ICP company, promotion with past evaluation, hiring surge with leadership change
- Moderate, single hiring signal, competitor engagement, topic post from a senior person
- Weak, single post engagement, headcount drift, new search match
Then apply a simple rule. One strong signal is worth acting on. Two moderate signals from independent causes are worth acting on. Weak signals need two others alongside them.
That is enough structure to improve a list meaningfully without pretending to a precision you cannot verify.
The timing conflict
Stacked signals usually have different decay windows, and this is where teams get it wrong.
A hiring surge is good for six weeks. A post engagement is dead in two. When both fire, the shorter window governs, you have the fortnight, not the six weeks.
Practically, this means a stacked signal should route to a faster lane than either signal alone. The combination is stronger and more urgent, which is counterintuitive and frequently handled backwards. More on the windows in signal decay.
Writing to a stacked signal
The temptation is to reference both. Do not.
“I saw you’re hiring three SDRs and that you engaged with [competitor]’s post about pipeline.”
That is two pieces of evidence that you have been watching closely, and it reads exactly as it sounds.
Reference one (usually the more public, less personal one) and let the other inform your angle:
Hi [name], saw you’re adding three SDRs this quarter. Most teams scaling outbound at that pace hit a deliverability wall around month two, before anyone notices the reply rate drifting.
Worth twenty minutes on what to set up before that happens?
The competitor engagement shaped the message (it told you they are evaluating, so the angle is “what to set up” rather than “what is outbound”) without appearing in it.
What stacking does not fix
A bad ICP. Two signals on a company you cannot sell to is two reasons to contact the wrong person.
A weak offer. Timing gets you read, not bought.
Poor execution. A perfectly timed message with broken personalisation still fails.
Stacking improves targeting precision. It does not substitute for knowing who you sell to or why they should care.
The short version
Single signals are noisy. Two independent signals pointing the same way are much less likely to be coincidence, and the hit rate reflects that.
Stack signals driven by different causes, hiring plus leadership change, promotion plus past evaluation, competitor engagement plus hiring in the relevant function. Do not stack a signal with your ICP filter, and do not count one signal observed three times as three signals.
Tier into strong, moderate and weak rather than building a weighted model you cannot validate. One strong or two moderate is the threshold.
When signals stack, act on the shorter clock. And reference only one of them in the message, however tempting the other is.
Next: how long each signal stays actionable, or scoring leads before enrolment.
Frequently asked questions
What is signal stacking in sales?
Acting on combinations of buying signals rather than single ones. Two independent signals pointing the same way are far less likely to be coincidental, which raises the hit rate of a prospect list considerably.
Which buying signals work well together?
Hiring surges with leadership changes, job changes into ICP-fit companies, competitor engagement alongside hiring in the relevant function, and promotions where the person previously evaluated you.
Should you mention multiple signals in outreach?
No. Referencing two pieces of evidence makes clear how closely you have been watching. Mention one, usually the most public, and let the other shape your angle.
How should stacked signals be scored?
Tier signals as strong, moderate or weak, then act on one strong or two independent moderate signals. Avoid weighted models unless you have the volume to validate the weights.