The short answer: A media sales ICP should describe more than company size and industry. Define five layers: firmographics, how the company buys media, who sits on the committee, which events make it ready now, and whether your team has a warm path in. Then tier accounts from it.
Most media sellers do not have a real ideal customer profile. They have a target account list, which is not the same thing. A target list is who you would like to sell to. An ICP explains why, and helps you decide which accounts get your limited time this quarter.
This guide walks through how to build one for media, ad tech, measurement, and data sales.
Why the classic ICP falls short in media
Standard software ICP frameworks emphasize firmographics and technology stack. Those matter, but they miss much of what shapes a media deal.
Media buyers do not buy because of company size alone. They buy because of a combination of size and category, how they buy media (in-house or through an agency), who sits on the committee, what is changing right now, and whether a credible person can introduce you. A useful media ICP has five layers, not two.
The five-layer ICP framework
- Firmographics: the baseline filters, such as ad spend range, vertical, geography, and growth stage.
- Buying model: agency-led, hybrid, in-house, or performance-only in-house, and which of these your product fits.
- Committee shape: the primary buyer, the economic buyer, the influencers, and the likely blocker.
- Triggering events: what has to be true right now, such as a new CMO, an agency change, or a launch.
- Warm paths: whether people on your team have real, recent relationships at the account.
Layer 1: Firmographics
Baseline filters: industry, geography, ad spend range, and growth stage. For illustration only, an enterprise inventory seller might set a floor around $10M in annual ad spend, while an ad tech product aimed at direct-to-consumer brands might start closer to $500K. These are example thresholds, not benchmarks. Set yours from your own closed-won data.
Layer 2: Buying model
Media buyers do not buy the same way. Segment by how they buy:
- Fully agency-led.
- Hybrid, with brand and agency sharing the work.
- In-house media team.
- Performance-only in-house team.
Your ICP should say which of these your product fits. If you sell to in-house teams, one useful clue is a cluster of in-house media hires at the brand over the last year or two.
Layer 3: Committee shape
Decisions involve many people. Forrester's summary of The State Of Business Buying, 2026 reports that, on average, 13 internal stakeholders and nine external participants influence a business buying decision. Your ICP should describe, by role:
- The primary buyer, by title and function.
- The economic buyer, such as a CMO, CFO, or VP of Media.
- The influencers, such as an analytics lead or agency planner.
- The likely blocker, such as procurement, legal, or an incumbent partner.
If you cannot describe those roles for a typical account, the ICP is not sharp enough. For the account-level version, see how to map the buying committee.
Layer 4: Triggering events
A working ICP says what has to be true now, not just in general:
- A new CMO or CFO in the last year.
- An agency of record change.
- A funding round, acquisition, or IPO.
- A product launch, category expansion, or new line.
- A renewal window with an incumbent partner.
- Public commentary on marketing efficiency or measurement.
An ICP without triggers is a static list. Our guide to buying signals covers which ones are worth acting on.
Layer 5: Warm paths
This is the most underused layer. For each target account, look at how much of your team's network already reaches it: direct relationships with the buyer, and ties to the buyer's manager, peers, former colleagues, and agency partners. Weight recency heavily. A conversation last quarter counts for far more than a connection from years ago.
Two accounts with identical firmographics are not equal if your team can get a credible introduction into one and not the other.
Account tiering
With the five layers in place, tier your accounts.
Many teams over-invest in Tier 3 because it is the biggest bucket. The tier list should move every month as triggers fire and relationships change. Our piece on account prioritization covers scoring in more detail.
ICP by seller type
Different media sales roles emphasize different layers:
- Inventory sellers (CTV, video, social): firmographics, buying model, triggers.
- Ad tech sellers: firmographics, committee shape, buying model, incumbent stack.
- Measurement sellers: committee shape, leadership changes, incumbent contract windows.
- Data and identity sellers: committee shape, privacy posture, clean room readiness.
If you sell several products, build separate ICPs. Inventory and measurement products often have different buyers inside the same brand. If you sell to agencies, build the agency ICP separately from the brand ICP.
A 30-day ICP sprint
- Week 1: Pull the last two years of closed-won and closed-lost deals. Look for shared firmographics, buying models, committee shapes, and triggers.
- Week 2: Interview the buyers behind five recent wins. Ask what started the conversation, who was involved, and who championed it internally.
- Week 3: Score your current account list on all five layers and identify Tier 1.
- Week 4: Rebuild outreach plans around Tier 1 first and Tier 2 second.
Rerun the sprint every six months.
A hypothetical example
The following is a hypothetical scenario for illustration. It does not describe a real customer or company.
A measurement startup sells to consumer brands. Its old target list had 300 brands chosen by revenue. After the sprint, the team learns that its wins came from brands with in-house analytics teams and a new marketing leader. Rescoring the list, only 40 accounts meet both conditions. Of those, 14 have a warm path through someone on the team. Those 14 become Tier 1, each with a named plan. The list got smaller and the plans got sharper.
Checklist: is your ICP real?
- It covers all five layers, not just firmographics.
- Thresholds come from your own wins, not borrowed benchmarks.
- It names the buying models your product fits.
- It describes committee roles, including the likely blocker.
- It lists the triggers that make an account ready now.
- It accounts for warm paths and their recency.
- It has a review date on the calendar.
Common mistakes
- Treating a target list as an ICP. A list is an output of the ICP, not the ICP.
- Building it once. Firmographics change slowly, triggers change weekly.
- Ignoring warm paths. It is often the best predictor of whether you get a first meeting.
- Segmenting only by industry. Buying model matters as much.
- Building it top-down. Start from your own deal data, then add judgment.
Quick answers
How many Tier 1 accounts should a rep have? Enough to give each one a real plan. For most reps that means a few dozen at most, not a few hundred. If a rep cannot name the next move for every Tier 1 account, the list is too long.
How often should the ICP change? Revisit firmographics and buying models about once a year, check triggers every week, and keep warm path information current continuously, because people change jobs all the time.
Can I score warm paths without a tool? For a small team and a short list, yes: ask each colleague to review the list and note who they know and when they last spoke. It gets impractical as the list and the team grow.
From one seller to another
An ICP is not a slide for the board deck. It is the thing that tells you which ten accounts deserve your Monday. Five layers, honest data, and a monthly tier refresh will do more than any new list.
If you want a step-by-step system for fit, timing, and access, the free Media Territory System course covers all three. Elo helps with layers four and five: inside Gmail and LinkedIn it shows dated signals at an account and who on your team has emailed, worked with, or is connected to people there, and how recently.
Sources: Forrester, The State Of Business Buying, 2026, blog summary published Jan 21, 2026 (checked Oct 24, 2026).
