The pitch is irresistible. Somewhere out there an account is researching your category, and if you could only see it, you could call at the exact moment they care.
The reality is a weekly list of forty companies who are apparently all in-market, most of whom are not, and a sales team who stopped opening it in week three.
Signal-based ABM does work. It works when you are ruthless about which signals you believe.
Signals we trust
A relevant job posting. A company advertising for a role that implies your problem is the strongest public signal available. It is specific, it is dated, it is verifiable, and it usually precedes a budget. A business hiring its first security engineer has decided something.
Funding, with a lag. A raise is real money and a public commitment to spend it. The mistake is calling the week it is announced, when everyone else is calling. Three months later, when the hiring has started and the honeymoon inbox has cleared, is better.
Leadership change in the function you sell to. A new head of anything reviews suppliers. This is the single best trigger for a displacement conversation and it is free to track.
Repeat visits to your pricing or implementation pages. Not one visit. A pattern, from more than one person at the same company, over a fortnight. In HubSpot this is a list rather than a tool purchase.
Technology change you can verify. A tool appearing or disappearing from their stack, when you can see it yourself rather than being told about it.
Signals we mostly ignore
Generic third-party intent scores. "This account is showing interest in your category" usually means somebody in a 4,000-person company read something. In a large organisation there is always someone reading something.
Single anonymous website visits. A visit from a company IP tells you very little. It is frequently a supplier, a candidate or a competitor.
Social engagement. Likes are not buying committees.
Anything you cannot trace to a person and a date. If you cannot say who did what and when, you cannot write an opening line that does not sound creepy or generic.
The rule that makes it work
A signal earns its cost when it changes what you say, rather than only when you say it.
If the only difference a signal makes is that the same sequence goes out sooner, you have bought an expensive timer. The point is that a funding round, a new VP and a job advert each justify a completely different opening, and the account can tell.
Which means the content has to exist before the signal fires. Most programmes get this backwards: they buy the data, then discover they have nothing specific to send.
How we set it up in HubSpot
You can run most of this without buying anything.
Create a custom property for signal type and signal date on the company record. Build active lists per signal, with recency conditions so a three-month-old trigger drops out. Use target account settings so the tiering is visible to sales in the record rather than in a spreadsheet.
Then the part everyone skips: put it in a task queue rather than a workflow. Signals should produce a human action with a named owner and a due date. Automating the response to a signal is how you get a sequence that says "I noticed you raised a Series B" nine weeks late.
Report on it at account level. Contacts known per account, engaged contacts per account, meetings from signal-sourced outreach. Lead counts tell you nothing here.
What it is worth
Done properly, on a list of fifty to a hundred accounts, this consistently produces conversations that cold outreach does not. It does not produce volume, and anyone promising both is selling.
The other half of the work is making sure that when a signal fires and someone looks you up, there is something credible for them to find. That is the ABM programme and the evidence behind it.
