Intent signals that actually mean budget, and the ones that waste your week

A signal is only useful if it tells you something has changed and roughly how long you have. Most of what gets sold as intent data fails that test.

Published 2026-09-01 by Deberli. About 9 minutes to read.

There is a version of go to market where you pick an ideal customer profile, build a list of everyone who fits it, and work through the list. It is the default because it is easy to explain to a board. It also produces the worst conversations in outbound, because fitting a profile is not a reason to talk to someone this week.

The alternative is to organise the week around change. Somebody just did something that makes your product relevant now, in a way it was not relevant a month ago. That is what a signal is, and most things sold as intent data are not one.

The test a signal has to pass

Two questions. Did something change, and do I know roughly how long it stays true.

Job title fails the first question. A VP of Sales was a VP of Sales last month and will be next month. It is a filter, and a necessary one, but it never tells you when to write.

A page view on your pricing page passes the first and fails the second in most setups, because by the time it reaches your CRM you often cannot tell whether it was yesterday or three weeks ago, and the difference is the entire value.

Anything that passes both is worth building a week around.

The ones that carry real urgency

Roughly in order of how much they tend to be worth, with the caveat that this ordering shifts by market and you should measure your own.

They are hiring for the problem you solve. The strongest signal available in public, because a job posting is a company committing real money to a problem, in writing, with a date on it. Two SDR roles means the pipeline is not repeatable. A first data hire means someone lost an argument about spreadsheets.
They are complaining about it in public. Rare, short lived, and close to unbeatable when you catch it. Someone writing about a workflow that is driving them mad has already done your qualification for you.
They asked for recommendations. The same thing with the buying intent stated outright. The window is days, because the replies to that post are your competition.
They raised a round or announced growth. New budget and pressure to spend it visibly. Weaker than it looks on its own, because everyone with a list buys the funding data and contacts them the same week.
They engaged with a competitor's post. Softer, but it tells you the topic is live for that person right now, which is more than a title ever will.
They engaged with your own posts. The warmest of the soft signals, and the one most teams forget to work at all.

How long each one stays warm

This is the part almost nobody plans around, and it decides whether the signal was worth having.

Asked for recommendations: days. Answer late and the decision is made.
Complained in public: a week or two, until they either solved it or gave up.
Hiring: the length of the search, often six to ten weeks. The most forgiving window on this list.
Raised a round: a quarter, though the first fortnight is crowded with everyone else who bought the same data.
Engagement with a post: about a week before the topic goes cold in that person's head.

Act inside the window or do not act on the signal at all. A stale trigger in an opening line reads worse than no personalisation, because it shows the data is old.

A signal you act on outside its window is just a cold email with a strange opening line. It reads worse than no personalisation at all, because it tells the reader you are working from stale data.

What this changes about the week

Signal led outbound produces a smaller list that has to be worked faster. That is an operational change, not just a targeting one.

A list built from a profile can sit in a spreadsheet for a month and lose nothing. A list built from signals decays daily, so the work has to be continuous: find today's, contact them today, and accept that the total volume will be lower than the profile approach would give you.

The trade is worth it because acceptance and reply rates move a long way, and because the volume you give up was mostly producing silence anyway.

Where the message has to change too

The signal is only half of it. If you find someone hiring two SDRs and then send the same message you send everyone, you have wasted the signal. The observation has to be in the first line, and it has to carry an inference the reader can agree or disagree with.

Not I saw you are hiring. That is a fact they already know. Instead: two SDR roles usually means the pipeline is coming off one or two people and it is not repeatable yet. Now they can nod or correct you, and either one is a conversation.

How we do it

Deberli watches eleven of these signals and sources people from them rather than from a static profile. The four groups are engagement with your posts or a competitor's, intent posts (hiring for the problem, complaining about it, asking for recommendations), funding and growth events, and the profile filters that narrow whatever those turn up. Recently changed jobs, follows a competitor and profile views are next.

The reason it is built this way is the same reason this article exists. Across 60 campaigns and 6,004 sourced prospects, 29 percent of invitations are accepted and 89 percent of replies come back positive or neutral. Those numbers are a targeting result before they are a copywriting one. A signal led list is a list of people who have a reason to answer, and that shows up in the reply rate long before it shows up anywhere else.

If you take one thing

Stop asking who fits and start asking who changed. The first question gives you a list that is always available and never urgent. The second gives you a shorter list with a clock on it, and the clock is what makes the message worth reading.