Turning Intent Data Into an ABM Tiering Workflow

This is about building an ABM account tiering workflow out of raw intent data: setting the signal thresholds, routing accounts into tiers, sequencing who touches the account and when, and auditing the whole thing once it stops predicting anything. That last part matters more than people admit, because every tiering model has an expiration date, and most teams find out the hard way. Let's start with the part nobody wants to hear. ### Intent Data Is a Rumor, Not a Verdict Raw intent data tells you someone at a company read something. Maybe they read a competitor's pricing page. Perhaps an intern was doing market research for a totally unrelated project. Maybe the CFO googled your category once and never again. You don't know. The data doesn't know either. It's a rumor with a timestamp. The mistake most teams make is treating intent signals as certainty: this account is in-market, full stop, send the SDR. But intent is closer to smoke than fire. Sometimes it means something's burning. Other times someone just left the toaster on. That's the whole game with intent data: it's an alarm system that can't tell the difference between an arsonist and a slice of burnt toast. I once watched a sales team scramble a full outbound blitz because a Fortune 500 account lit up on a competitor comparison topic for three straight days. Everyone assumed a buying committee was in motion. It turned out to be one procurement analyst doing a mandatory annual vendor review, a task about as urgent as flossing. The account didn't buy for another fourteen months. The lesson stuck around a lot longer than the deal did. So before you build a tiering model, accept the premise: you're measuring the probability of buying intent, filtered through a vendor's tracking pixel and a topic taxonomy that may or may not match how your buyers actually talk about their problems. ### Setting Signal Thresholds (Or: How Not to Cry Wolf) Thresholds are where most ABM programs quietly fall apart. Set the bar too low, and every account in your TAM lights up like a slot machine. Set it too high, and you miss the accounts actually circling the drain toward a purchase decision. Here's the practical approach: - Start with a composite score. One intent spike from one keyword on one vendor's platform isn't a threshold. It's noise wearing a costume. Combine firmographic fit, intent surges across multiple topics, and engagement with your own site or content. Three weak signals stacked together beat one strong signal alone. - Calibrate against your own closed-won data. Every intent provider ships with an out-of-the-box scoring model, but those defaults are a starting point at best. Pull your last several quarters of closed deals, look at what intent activity actually preceded them, and set your threshold there. If your best customers showed intent spikes 60 days before they entered a sales conversation, that's your window, more useful than the 14-day default someone configured in a demo. - Build in a decay function. Intent signals rot. An account that spiked hard four months ago and has gone silent since carries a different weight than one that spiked yesterday. If your platform doesn't decay scores automatically, do it manually in your CRM logic, because static scores treat a warm lead from last spring like it's still warm today. The threshold isn't a science experiment you run once. It's a knob you keep turning, and if you're not adjusting it quarterly, you're managing a dashboard by squinting at it and hoping, not managing a model. ### Routing Accounts to Tiers Once you've got a score you trust, tiering is really just triage. Think of it like an ER waiting room: not everyone gets the same doctor at the same speed, and that's not unfair, that's math. A standard three-tier structure looks something like this, though the exact labels matter less than the discipline behind them: - Tier 1, the small number of accounts getting the full-court press: personalized outreach, executive sponsorship, custom content, the works. - Tier 2, a mid-size bucket getting targeted but scaled plays: sequence emails with some personalization, retargeting ads, maybe a webinar invite. - Tier 3, everyone else, running on largely automated nurture until they earn their way up. The routing logic has to be boring and mechanical, or your reps will find ways to argue every account into Tier 1. (They will do this. Reps believe every account is Tier 1 the same way every parent believes their kid is gifted.) Set the score bands, automate the routing in your CRM or ABM platform, and remove humans from the initial assignment entirely. Humans get a vote later, during review, but that vote belongs after triage, not during it. Q: How many sales reps does it take to admit an account belongs in Tier 3? A: None. It's always somehow a Tier 1 in disguise. ### Sequencing the Touches Tiering tells you who gets attention. Sequencing tells you in what order, through which channel, and who's actually responsible for the next move. For Tier 1, the sequence usually looks like marketing warming the account with targeted ads and a piece of custom content, followed within days by an SDR or AE reaching out with something specific to that account's intent signal, distinct from a generic "checking in." If the account was spiking on a competitor comparison topic, the outreach should reference that, directly and without pretending it's a coincidence. For Tier 2, the sequence stretches out. Marketing carries more of the weight, sales touches come in at defined intervals rather than immediately, and the personalization gets lighter but not absent. Tier 3 runs almost entirely on automation until a signal promotes the account upward, at which point the sequence for the new tier kicks in. The part teams skip: defining what happens when an account moves tiers mid-sequence. If a Tier 3 account suddenly spikes hard enough to qualify for Tier 1, does the nurture sequence stop immediately? Does sales get an alert within the hour or find out next Monday? Build the tier-change trigger into your workflow before you launch, not after your best rep complains they found out three weeks late. ### When Tiers Stop Predicting Anything Here's the part that separates a real ABM program from a very expensive spreadsheet: eventually, your tiers will stop working, and you have to notice. Signs it's happening: - Tier 1 accounts close at roughly the same rate as Tier 2 accounts. If your top tier isn't converting meaningfully better than the tier below it, the model has stopped tiering anything and started merely relabeling your pipeline. - Sales stops trusting the tier assignments and starts working accounts off gut feel instead. This is the loudest signal of all, because reps vote with their calendars. When they're ignoring the tier list, they've already run the audit for you, informally, and the model failed. - The intent topics you calibrated against eighteen months ago don't match how the market talks now. Buyer language drifts. Vendor names get folded into new categories. A taxonomy built two product cycles ago is measuring a market that's moved on. The audit itself doesn't need to be complicated. Pull closed-won and closed-lost deals from the last two or three quarters, map them back to the tier they were assigned at the time, and check whether tier actually correlated with outcome. Rebuild the thresholds from scratch using fresh data if it didn't, the way you did the first time, rather than patching the old ones. Bolting a fix onto a broken model is like re-taping a leaking pipe for the fourth time: at some point you're not fixing the pipe, you're just delaying the flood. Run this audit on a calendar, not when someone finally complains loud enough. Quarterly is reasonable for most B2B sales cycles. Skip it, and you end up exactly where most ABM programs end up two years in: a tiering model nobody trusts, dashboards nobody opens, and a sales team quietly running its own spreadsheet in the background because the "official" one lost the room a long time ago.