ICP framework for tech consultancies

An ICP is a filter, not a description

The most common mistake is treating the ICP as a marketing artifact: a paragraph on a slide that describes an aspirational customer. A useful ICP is a set of binary filters that any sales rep can apply in 30 seconds against an account list. If your ICP cannot be used to disqualify, it is not an ICP. It is a wish.

The five-filter framework

1. Vendor alignment: which technology stack must be present for your delivery to be valuable? Be specific (Snowflake + dbt, not 'modern data stack').
2. Company size band: revenue or headcount range where your engagement model fits. Both ends matter; too small and budgets are absent, too large and procurement kills momentum.
3. Trigger event: the operational pain that creates urgency. Migration, audit failure, leadership change, funding round, new compliance requirement.
4. Buying authority: which role can write a check at the size of your typical engagement without escalation. If that role is absent, the account is not ready.
5. Cultural fit: signals that the buyer values quality of advice over lowest price. Existing investment in tooling, presence at industry events, public technical content.

Test the filter against your actual book

Take your last 10 closed-won engagements and your last 10 closed-lost. If your filter does not correctly classify at least 8 of the 10 wins as ICP and at least 8 of the 10 losses as non-ICP, the filter is wrong. Refine until it does. This exercise is uncomfortable because most firms discover their best clients do not match the ICP they have been telling the market about.

Use the filter everywhere, not just in sales

An effective ICP filter changes content (you write for that persona, not for everyone), changes outbound (lists are filtered before any message is drafted), changes hiring (sellers are evaluated against ICP knowledge), and changes pricing (you stop discounting for non-ICP accounts that will never compound). The compounding effect of consistent ICP application is the single largest GTM lever a consultancy has.

Revisit annually, not quarterly

ICPs should be stable enough that the market notices. Changing it every quarter signals a firm that does not know what it is. Revisit formally once a year, after analyzing the prior 12 months of wins, losses and gross margin. Adjust at the edges. Wholesale changes are almost always a sign of a different problem (positioning, pricing, or delivery quality) being misdiagnosed as a targeting problem. Once the filter is stable, apply it across every vendor playbook: SAP, Microsoft and AWS.