---
title: "Account tiering for outbound: research budget by tier"
description: "How to tier accounts by potential value and set a cost-per-account research budget for each tier, with real thresholds."
date: "2026-08-29"
tags: "prospecting, account based selling, outbound strategy, sales ops"
readTime: "16 min read"
slug: "account-tiering-outbound"
canonical: "https://firstsales.io/blog/account-tiering-outbound/"
---

# Account tiering for outbound: research budget by tier

**TL;DR:** Most teams research every account the same amount, which means they underinvest in the accounts that matter and overinvest in the ones that never had a chance. Tier accounts by realistic deal value and buying readiness, then set a research budget in minutes per account for each tier. A tier 1 account might earn 20-30 minutes of human or AI research. A tier 3 account should get 2-3 minutes, or none.

---

**Table of contents**
- [Why flat research spend is the default mistake](#why-flat-research-spend-is-the-default-mistake)
- [What tiering actually means](#what-tiering-actually-means)
- [Building the tier criteria](#building-the-tier-criteria)
- [The cost-per-account budget model](#the-cost-per-account-budget-model)
- [What research buys at each tier](#what-research-buys-at-each-tier)
- [Where AI changes the math](#where-ai-changes-the-math)
- [A tiering workflow you can run this week](#a-tiering-workflow-you-can-run-this-week)
- [Common tiering mistakes](#common-tiering-mistakes)
- [Comparison: research effort by tier](#comparison-research-effort-by-tier)
- [FAQ](#faq)
- [Conclusion](#conclusion)

---

Every outbound team says they focus on their best accounts.

Almost none of them actually spend their research time that way.

Pull up a campaign in most tools and you will find the same five data points pulled for every account on the list: company name, industry, headcount, a generic pain point, and whatever the enrichment vendor happened to return.

That is not tiering. That is a flat rate applied to accounts with wildly different potential.

A 2,000-employee account that fits your ideal customer profile perfectly deserves ten times the research spend of a 40-person company that barely clears your minimum deal size.

Most teams give them the same three minutes.

## Why flat research spend is the default mistake

Flat research spend happens because tiering feels like extra work up front, and most outbound tools do not force you to declare a tier before you start sending.

You pull a list, run it through enrichment, and start drafting.

The list has no structure, so the workflow has no structure either.

The cost shows up later, in two directions at once.

On the high-value accounts, you send a generically personalized email to a buyer who could be a $150,000 deal, using the same personalization depth you used on a $4,000 deal three rows up.

That account gets one shot at a first impression and you spent it on autopilot.

On the low-value accounts, you burn research minutes, sending capacity, and follow-up cadence on companies that were never going to close, or that would close for so little that the acquisition cost eats the margin.

Neither side is a research problem exactly. It is a budgeting problem that outbound teams do not usually treat as a budget.

Sales teams that fully replaced human SDRs with AI made up about 22% of the market by 2026, with another 45% running a hybrid model.

Only around 2% of the full-replacement attempts actually stuck.

The teams that stuck were disciplined about where the AI spent its time and where a human still had to look.

Tiering is the mechanism that makes that discipline possible.

## What tiering actually means

Account tiering sorts your target list into groups based on realistic deal value and buying readiness, then assigns a different research and outreach investment to each group.

It is not the same thing as your [ideal customer profile](/blog/ideal-customer-profile).

ICP tells you who belongs on the list at all.

Tiering tells you how much attention each account on that list deserves once it has qualified for inclusion.

A common structure looks like this.

Tier 1 accounts are your best-fit, highest-value targets, usually 5-15% of a list.

Tier 2 accounts are solid fits with moderate value, usually 30-40% of a list.

Tier 3 accounts are everything else that still clears your minimum bar, usually the remaining 50%+.

Some teams add a tier 0 for named strategic accounts, a small handful, sometimes fewer than 20, that get account-based treatment closer to enterprise sales than to outbound at scale.

The tiers are not about company size alone.

A 500-person company in a segment where you have never closed a deal might rank below a 60-person company that looks exactly like your last five wins.

## Building the tier criteria

Tier criteria should combine fit and readiness, not just firmographic size.

Fit criteria describe how closely an account resembles your best customers: industry, tech stack, team size, growth stage, and any [compound buying signal](/blog/compound-buying-signals) that correlates with past wins.

Readiness criteria describe whether this specific account is likely to be receptive right now: a recent [hiring signal](/blog/hiring-signal-outbound), a job change at the buyer role, funding activity, or a fiscal year timing window opening up.

These signals fire and fade quickly, which is why a readiness score needs a refresh cadence, not a one-time calculation done at list-build time and never touched again.

A simple scoring approach works better than an elaborate model most teams will not maintain.

Score fit from 0-10 based on how many ICP criteria the account matches.

Score readiness from 0-10 based on how many active signals are present.

Add them, and set tier cutoffs on the combined score: 15+ is tier 1, 8-14 is tier 2, below 8 is tier 3 or excluded.

This is intentionally rough. A scoring model that takes two weeks to build and requires a data scientist to maintain will get abandoned by month three.

A spreadsheet with five weighted columns that a sales ops person updates monthly will survive.

```mermaid
graph TD
    A[Full target list] --> B{Fit score}
    B -->|High fit| C{Readiness score}
    B -->|Low fit| X[Excluded or backlog]
    C -->|High readiness| D[Tier 1: full research]
    C -->|Moderate readiness| E[Tier 2: standard research]
    B -->|Moderate fit| F{Readiness score}
    F -->|Any| G[Tier 2 or Tier 3]
    G --> H[Tier 3: light research]
```

## The cost-per-account budget model

Once accounts are tiered, assign each tier a research time budget stated in minutes, not a vague instruction to "personalize more."

Minutes are the unit that actually gets enforced, because they translate directly into headcount, AI spend, or both.

A workable starting model:

Tier 1: 20-30 minutes per account. This covers reading the company's recent public activity, checking the buying committee on LinkedIn, reviewing a specific trigger event, and drafting a first message that references something true and current about the account.

Tier 2: 5-10 minutes per account. Enough to confirm the fit signal, pull one specific detail, and personalize the opening line without a deep research pass.

Tier 3: 1-3 minutes per account, largely automated. A template with light variable substitution, sent at volume, with no manual research step.

Multiply minutes by your fully loaded hourly cost for the person or system doing the research, and you get a cost-per-account figure you can actually defend in a budget conversation.

At $40 an hour loaded cost for an SDR, a tier 1 account at 25 minutes costs about $17 in research alone, before any tooling or list cost.

A tier 3 account at 2 minutes costs about $1.30.

That thirteen-times gap is the point. It should exist, and right now, for most teams, it does not.

## What research buys at each tier

The budget only matters if it buys something different at each level, not just less of the same thing.

Tier 1 research should produce a specific, non-generic reason the account is being contacted now, ideally tied to a real trigger rather than a firmographic match.

It should also map the buying committee, since [multithreading into a buying committee](/blog/multithreading-outbound-buying-committee) matters most on the deals big enough to have one.

Tier 2 research should confirm one fact that makes the email feel researched rather than templated, without the full committee mapping or trigger-hunting that tier 1 gets.

Tier 3 outreach should skip individualized research and instead rely on segment-level personalization: a shared vertical pain point, a shared tech stack detail pulled from enrichment data, applied consistently across the whole tier.

This is where AI-assisted research earns its keep, because a model can pull and summarize public signals for a tier 3 batch in seconds per account, at a cost per account that a human doing the same task could never match.

That does not mean tier 3 should get zero attention. It means tier 3 attention should scale through automation instead of headcount, which is a different design decision than skipping it outright.

FirstSales runs exactly this kind of tiered research inside a single workflow: heavier signal-based research and drafting time on higher tiers, lighter automated passes on the rest, with a human checkpoint before anything sends.

That checkpoint matters more as the tier gets higher, since a bad tier 1 email costs you a real opportunity, while a bad tier 3 email costs you one send out of a batch.

![Account tiers arranged by research minutes and deal value](/images/blog/account-tiering-outbound/inline-1.webp)

## Where AI changes the math

AI research tools did not eliminate the tiering decision. They changed what a given research budget buys.

Five years ago, 20 minutes of research meant a human reading a LinkedIn profile, a company blog, and maybe a news search.

Today, the same 20 minutes, mostly spent reviewing an AI-drafted brief rather than compiling one from scratch, can cover a wider set of sources: recent hires, funding events, product launches, and executive commentary, summarized and cross-checked.

That shift pushes some accounts that used to sit in tier 2 up into effective tier 1 treatment, because the marginal cost of deeper research dropped.

It also means tier 3 no longer has to mean "no personalization at all."

A model can pull one real, current fact per account even at high volume, which was not economical with a human doing the same task at the same volume.

The risk is treating this as permission to skip tiering altogether because "AI does the research now."

AI-supported human SDR teams built 2.8x more pipeline than manual-only teams, but that gap came from AI freeing up human time for the accounts that needed it, not from applying AI evenly everywhere.

Teams that point the same AI research depth at every account regardless of tier end up back at flat spending, just automated.

The tiering decision still has to happen before the AI does anything. AI changes the cost curve inside each tier. It does not remove the need to draw the tiers.

## A tiering workflow you can run this week

Start by pulling your list of closed-won deals from the last 12-18 months and tagging the firmographic and behavioral traits they share.

This becomes your fit-scoring rubric, built from evidence rather than assumption.

Next, define 3-5 readiness signals you can reliably detect, whether through a data provider, a [waterfall enrichment](/blog/waterfall-enrichment-b2b-data) process, or manual review, and weight them.

Run your current target list through both scores and sort into tiers using the cutoffs above as a starting point, adjusting the split once you see how many accounts land in each bucket.

Assign a research minutes budget to each tier and write it down somewhere your team actually references, not buried in a slide from a kickoff meeting.

Then run one full cycle and compare reply rate and meeting rate by tier.

If tier 1 is not meaningfully outperforming tier 3 on reply rate, your tier criteria are wrong, not your research depth.

If tier 3 is matching tier 1 on reply rate, you are overspending on tier 1 research relative to what it is buying you.

Revisit the tier cutoffs quarterly. A segment that validated well, the kind of validation covered in [testing a segment before scaling](/blog/segment-validation-test-outbound), can move from tier 3 to tier 2 as evidence accumulates.

## Common tiering mistakes

The most common mistake is building tiers around company size alone, ignoring fit and readiness entirely.

A large company that is a poor fit for your product is not a tier 1 account. It is a bad account that happens to be big.

The second mistake is setting the research budget and never enforcing it, so tier 3 accounts quietly absorb 15 minutes of manual attention because a rep likes the logo.

The third mistake is tiering once and never revisiting it, treating the tier list as a static artifact instead of something that should move as signals change and deals close or die.

The fourth mistake, and the one that shows up most in AI-assisted outbound specifically, is applying the same AI prompt and research depth across every tier because it is easier to build one workflow than three.

That collapses the entire point of tiering back into a flat rate, just automated instead of manual.

## Comparison: research effort by tier

| Tier | Share of list | Research minutes | Depth of research | Sending approach |
|---|---|---|---|---|
| Tier 1 | 5-15% | 20-30 min | ✓ Trigger-specific, buying committee mapped | ✓ Manual review before send |
| Tier 2 | 30-40% | 5-10 min | ✓ One confirmed fact, no committee mapping | ✓ Spot-checked before send |
| Tier 3 | 50%+ | 1-3 min | ✗ No individual research, segment-level only | ✓ Automated with pre-send checks |
| Flat spend (no tiering) | 100% | Same for all | ✗ Generic depth everywhere | ✗ No differentiated review |

## FAQ

### What is account tiering in outbound sales?

Account tiering is sorting your target account list into groups by fit and buying readiness, then assigning each group a different level of research and personalization before outreach.

### How many tiers should an outbound program use?

Three tiers work for most teams. A fourth tier for a small set of named strategic accounts makes sense once you have more than a handful of accounts big enough to warrant account-based treatment.

### What percentage of accounts should be tier 1?

Most programs land tier 1 at 5-15% of the full list. If more than a quarter of your list qualifies as tier 1, your criteria are too loose to be useful.

### How much time should I spend researching a tier 1 account?

A common range is 20-30 minutes per account, covering a specific current trigger, a buying committee check, and a first message built on both.

### How much time should tier 3 accounts get?

1-3 minutes, usually through automated or template-based personalization rather than individual manual research.

### Does account tiering apply to inbound leads too?

Yes. The same fit and readiness logic works for scoring inbound leads, though inbound already carries an intent signal that outbound accounts have to earn through other signals.

### How is tiering different from lead scoring?

Lead scoring usually happens at the individual contact level and often includes engagement behavior. Account tiering happens at the company level before any outreach starts, based on fit and external signals.

### Can AI do the tiering itself?

AI can help score fit and detect readiness signals at scale, but the criteria and cutoffs should come from your own closed-won data, not a generic model.

### What if I do not have enough closed-won data to build fit criteria?

Use your ICP definition as a starting proxy, run one cycle, and refine the criteria once you have real win and loss data from that cycle.

### Should tier assignments change over time?

Yes. A signal firing today can move an account from tier 3 to tier 1 temporarily, and accounts should be reviewed at least quarterly against updated closed-won patterns.

### Does tiering slow down outbound velocity?

It changes where velocity goes. Tier 3 moves faster because it skips manual research. Tier 1 moves slower per account but should convert at a meaningfully higher rate.

### What tools help with tiering?

A CRM field for tier, a scoring spreadsheet or simple model, and a research or drafting tool that can apply different depth by tier are the minimum. Complex tiering software is not required to start.

### How do I know my tier cutoffs are wrong?

Compare reply rate and meeting rate across tiers after a full cycle. If tier 3 performs close to tier 1, your criteria are not discriminating between good and bad accounts.

### Does account tiering work for small TAMs?

Yes, though the tier sizes shrink. See our [small TAM playbook](/blog/small-tam-outbound-playbook) for how tiering changes when your whole addressable market is a few hundred accounts.

### Should tier 1 accounts always go through a human before sending?

Generally yes. The cost of a bad email to a high-value account is much higher than the cost of the extra review time it takes a person to catch a mistake before it ships.

### How does tiering interact with multithreading?

Multithreading a buying committee is expensive per account, so it should be reserved for tier 1 and possibly tier 2, not applied across an entire list.

### Can a tier 3 account become tier 1 later in the relationship?

Yes, especially if it responds, engages, or a new signal fires. Treat tiers as a snapshot that updates, not a permanent label.

### What is the biggest sign a team is not really tiering?

The research depth and message quality look identical across every account in the CRM, regardless of stated tier.

### Does tiering reduce total outreach volume?

Not necessarily. It reallocates where the volume comes from, usually keeping or increasing tier 3 volume while cutting sloppy volume in what should have been tier 1.

### How often should tier criteria be rebuilt from scratch?

Full rebuilds are rarely needed. Incremental weight adjustments each quarter, based on new closed-won and closed-lost data, keep the model current without constant rework.

## Conclusion

Tiering is not a nice-to-have layer on top of outbound. It is the budget allocation decision that determines whether your best accounts get the attention they are worth.

Flat research spend feels fair because every account gets the same treatment. It is actually the opposite of fair, since it shortchanges your best opportunities to subsidize accounts that were never going to close.

Set the tiers from your own closed-won data, not a generic framework. Assign minutes, not vibes, to each tier.

Then measure whether the tiers actually predict reply rate, and adjust the cutoffs when they do not.

The gap between a well-tiered program and a flat one shows up first in reply rate on your best accounts, and eventually in how much pipeline your team can carry without adding headcount.