Act on early signals without flooding reps: map before you pitch

Which signals predict a buying cycle 6–12 months out?

The signals that predict a cycle that far out change a seat or a system the buyer depends on. Watch for a new leader in a seat that buys, hiring for the team that would run your product, an announced migration, a funded project and the current vendor’s contract nearing its notice date. A new CIO, a RevOps hire and a CRM migration all qualify. Whether they predict your own deals is a question your CRM can answer.

Run the check backwards from the deals you already have. Take your last 20 RFPs and won deals, 10 at the least, and list the dated events at each account in the 12 months before the RFP or the first meeting. Keep the 3–8 signal types that came before most of them, with the median months from signal to RFP. Drop the rest.^↗

The timing shows why the RFP feels late. In 6sense’s 2025 survey of nearly 4,000 buyers, the average buying cycle ran 10.1 months.^↗ Buyers first contacted a seller 61% of the way through it.^↗ That is about 6 months in.^↗ By then, 94% of buying groups had ranked their shortlist, and the winner was on the Day One shortlist 95% of the time.^↗ When the RFP is the first you hear of a deal, you’re reading the last step of a decision that started months earlier.

Lists of signals are easy to find; what to do next is not. All 4 pages that AI answers cite for this question list signal types. None turns a signal into a mapping task on the buying group, limits what reaches a rep or checks whether a seat was engaged before the RFP.^↗

Early signals should start a map, not a pitch

We think a structural signal needs an owner and an action the same day, and that the action should be mapping the buying group. ChatGPT’s answer to this question in October 2026 points the other way: watch and educate accounts 6–12 months out with no rep task, and map the stakeholders 3–6 months out.^↗ Waiting, pitching and alerting every rep each lose the signal in a different way.

Waiting misses the seat that changed, which is where the plan gets written. A new CIO or a new head of RevOps sets the agenda the requirements later come from. An educational email doesn’t engage that person. A single good question, answered, does.

Pitching is early by definition. A signal 6–12 months out arrives before the account is in market, and at any given time 95% of potential buyers aren’t.^↗ We think a pitch then spends the new seat holder’s first impression of you on a conversation they can’t have yet.

Alerting every rep gets the signal ignored. Sales teams already leave about 70% of marketing leads unpursued.^↗ We expect a raw feed of signal alerts to go the same way.

In our view the gap is usually routing and process; the data is rarely what’s missing. Most teams already see these signals in a news feed or an intent tool. What they lack is a named owner, a defined action and a date. The map also designs multi-threading in instead of leaving it to the champion: it names who talks to whom and by when. None of this means a signal predicts a purchase on its own. It’s a start line.

Route every signal to a lane the same day

Every signal gets a lane on the day it arrives: map, sell or log. A structural signal at an account that passes your fit gate starts a mapping play. Evidence of an active evaluation goes to the account’s AE as a sales alert. Everything else is logged on the account, where it moves the score and waits for the next review.^↗

Add 2 rules so the lanes don’t flood sales. First, keep a single open signal task per account: a second signal at the same account joins it instead of creating another. Second, cap the mapping plays each AE owns per week at what they can work beside their open deals. Overflow goes to a named SDR who covers that tier. Leave the sales lane uncapped: evaluation evidence is rare, and it goes cold fastest.^↗

The mapping play is short. It keeps the clock of our early-signal routing rule: an owner the same working day, a first human touch within 1 working day and the owner’s manager after 2 working days without one.^↗

When evaluation evidence arrives months later, the sales alert lands at an account where a seat already knows you.

How many signal tasks should 2,000 accounts send a rep?

A handful a week at most: in this worked example, each AE gets 1–2.^↗ Take a fictional vendor of configure-price-quote (CPQ) software that sells €80k-a-year deals to 2,000 named companies, with 6 AEs and 2 SDRs. Tier 1 holds 60 accounts, 10 per AE; Tier 2 holds 240 and Tier 3 the other 1,700.^↗ At €80k our starting point is 4 required seats: RevOps as the champion, the sales leader who holds the budget, IT for the CRM integration and finance for pricing approval.^↗

In a typical week its sources log 44 signals at 30 accounts:^↗

• 6 accounts fail the fit gate: logged
• 5 show only an intent surge: logged, and their scores rise
• 3 show evaluation evidence (a request for information, a consultant’s vendor survey and a reference request): 3 sales alerts the same day
• 16 show a structural signal: 16 mapping plays, 4 at Tier 1 accounts for their AEs and 12 for the SDRs, among them 5 Tier 3 accounts that move up a tier the day the signal lands

Each AE gets 1–2 signal tasks that week, against about 7 if every signal went to the AEs. Each SDR runs 6 mapping plays, about 3 hours at 30 minutes each. The vendor’s cap of 3 mapping plays per AE a week isn’t reached.^↗

At 90 days, 9 of the 16 plays close as engaged, 4 as named only and 3 as missed. All 3 misses are job posts for roles that never touch quoting, so the vendor narrows that signal to the roles that do. Over the year, 12 RFPs arrive from the 2,000 accounts. At 7, a seat is engaged before the RFP goes out; at 3, a signal is seen but no seat engaged; at 2, nothing is seen. Both of those follow a new CFO, so a change of finance leader joins the library.^↗

Record the seat a signal opens, then count the misses

Record the seat and the outcome on the signal itself, in the CRM, or you can’t tell which signals work. These are the fields we put on every routed signal:^↗

Then measure 2 things to see whether the lanes work. The first is per signal: at 90 days, did the play end with an engaged seat, only a name or a miss? Engaged means a two-way exchange with the seat’s holder, such as a call, a meeting or a written reply to a question you asked. Report it monthly by signal type, and redefine or drop the types that rarely end engaged.^↗

The second starts from each RFP and looks back. For every RFP or new opportunity at a target account, record whether a signal came before it and whether a required seat was engaged before it went out. Each is a hit (a seat engaged), a routing miss (a signal seen, no seat engaged) or a library miss (nothing seen). Track the share of RFPs with a seat engaged before release each quarter, against the same count for the year before, rebuilt from the CRM.^↗

Watch the reps’ load beside it. Check S4 in the Panel Check, our GTM audit, scores the share of alerts followed by an activity within 7 days. Its reason: “Alerts without an owner become noise within a month.”^↗ If that share falls as signal volume rises, tighten the fit gate before you add signal types.