Try this: Ask stakeholders at your company to define your ICP.

Every answer will probably be different.

Each person – from the Head of Marketing to the CEO to Head of Sales – will give you an answer that reflects the data that matters to them. These may seem like small differences but they actually have a big impact on your business.

Often the answers will also include qualitative attributes that aren’t actionable or measurable. Think of “data-driven” or “tech-savvy.” These are common descriptions that come from research done by the product marketing team, but can’t be converted into useful targeting criteria.

Or the description might include attributes that seem to be targetable but are not. Describing your ICP’s industry as “B2B SaaS” is a good example of this. You’d assume B2B SaaS to be a commonly-used classification, but it’s not.  

The problem with how ICPs are created today

Most companies start defining their ICP with a bit of intuition—i.e. a founder has a prediction about their perfect customer or there’s some early customer feedback. Then, the ICP is solidified through additional market feedback, user research, and analysis of GTM data.

Over time, teams will start to make tweaks to the ICP definition based on the data that’s important to them and the systems they’re operating in. For example, marketing will optimize their ICP definition to find more leads inside Linkedin, while sales will think about who makes a good “sales-qualified opportunity” in Zoominfo.

When the process isn’t systematic or data-informed, the result is that key teams start speaking different dialects. Eventually, conversion starts to fall because sales, marketing, and product are all moving in different directions. Marketing will bring in leads that don’t convert, while sales may convert opportunities that aren’t served in the best way by the product.

The bottom line is that a data-driven view of your ICP that everyone can buy into is critical. But how can you get there? Here are the steps I, as Primer's Head of Sales, recommend:

Step 1: Spend time together understanding the objective of building your ICP

Identify each team’s use case for an ICP. For example, marketing wants an ICP definition to help with targeting criteria while the product team wants it in order to create a product roadmap. Figure out how to create a central ICP definition that can extend into each of those use cases.

Step 2: Gather inputs from Sales, Marketing, Product – and CS.

Often, customer success (CS) teams are ignored in the data gathering process. But asking them for their most successful customers will likely lead you to those who renew. Focusing solely on marketing or sales can lead to leads that don’t close, or closed deals that don’t succeed.

By following the entire funnel as you identify your ICP, you’ll gather valuable data points on what matters to creating a successful customer from the top of your funnel all the way through it.

Step 3: Identify input data that could potentially skew your results.

Weigh all your input data based on identified gaps in the funnel, and then discount it accordingly. For example, you could be seeing a drop off between lead-to-opportunity, which would lead you to identify that some 30% of the leads you’re getting aren’t a great fit for your business. You’d then have to reduce the weightage of that data in the model you’re building.

Step 4: Enrich with additional data points.

Next, enrich your primary dataset with as many additional data points as possible from an enrichment dataset like Apollo, Zoominfo, or Clearbit. We have a list of recommended filters/attributes at the bottom of this post. At Primer, we maximize the amount of information we can collect by using a combination of several databases.

Once you have an enriched dataset, you’ll analyze it to find the points of correlation. Then you’ll decide how tightly you want to define your ICP. For example, when we conducted this ICP analysis at Primer, we found that there was an 80% correlation between sales-qualified opportunity and an employee count of 11-2K (i.e. 80% of our opportunities came from companies with 11 to 2,000 employees.) We could’ve refined that further to a 90% correlation, and narrowed our ICP profile to companies with 50 to 2,000 employees but we decided to keep it broad.

Step 5: Get buy-in from the entire company

Take your analysis to all stakeholders at the company (Marketing, CS, Sales, Product) and ensure that there is wide alignment on your newly-defined ICP. Ask them if they want to include any additional data sets and make sure that they’ll be able to use the definition in their own work.

Examples of filters/attributes

Company-level attributes:

• Location
• Employee count
• Industry
• Revenue
• Funding
• Website traffic
• Technology stack

Person-level attributes:

• Title
• Department
• Seniority