Smaller companies tend to know their ideal customer profile (ICP) well. They have to. They’re still ironing out product issues and developing features, making selling to a broad audience tricky.
Unfortunately, just because you know your ideal target market doesn’t mean you can use these data points in your ICP model. We’ve seen many organizations develop a highly detailed scoring model with gating mechanisms only to shoot themselves in the foot.
Before you build out your ICP scoring, it’s essential to understand your data.
What Is an Ideal Customer Profile?
An ideal customer profile is the combination of firmographic and demographic attributes that can be used to predict whether or not someone is likely to purchase your product. Sometimes “ICP” is used interchangeably with very different terms–and it shouldn’t be.
The three categories are:
- Target Account: the ideal firmographics that align with your product (example: organizations with over 1,000 employees in the technology industry)
- Buyer Persona: the ideal demographics of people who purchase your product (example: people in the information technology department with a role of manager or greater)
- Ideal Customer Profile: a combination of firmographics and demographics that signal the ideal human target at the right company (example: a person in an information technology department with a role of manager or higher that is employed at an organization with more than 1,000 employees in the technology industry)
Because we’re ultimately selling to people and organizations in B2B marketing, it’s important to understand which traits (both human and organizational) correlate to a higher propensity to purchase.
Why Is an Ideal Customer Profile Useful?
Ideal customer profiles can be used to make your outbound marketing campaigns more effective and help your sales team quickly target the most relevant leads.
For example, a person with the right firmographics and demographics should be contacted by sales the minute they fill out a Contact Us form. If a person has the right firmographics but wrong demographics, it’s still a good idea to reach out to them ASAP. Suppose the person is a good fit but they work for a company that doesn’t have the means to purchase your product. In that case, you’ll want to keep a relationship just in case the company is wildly successful, or the person moves to a new, larger organization. If it’s a poor fit all the way around, they’ll go to the bottom of the pile.
Using ICP logic, you can help the sales team quickly prioritize their inbound and outbound prospecting. Let’s face it, they’re already doing this exercise on their own. Isn’t it better to make sure they’re working off data that actually correlates to a win? Humans are biased beings and are hardwired to remember negative experiences more readily than a positive experience. Don’t let your sales team throw out excellent leads because of one too many negative experiences.
Similarly, understanding your contact database helps marketers better prioritize their campaign efforts. If you only have a few thousand dollars to spend on a direct mail campaign, with ICP in place you can only send it to the people who have the right firmographic and demographic traits. ICP can also inform paid social advertising or email campaigns and determine which trade shows make the most sense to attend.
How to Create an Ideal Customer Profile
While it may be tempting to take your sales executive’s definition of the ICP at face value, a lot more goes into designing a reliable ideal customer profile model.
01 Choose a Point of Optimization
If your focus is increasing new customer acquisition and you have more than 50 Closed Won opportunities, you likely have enough data to start compiling a model. If you have less than fifty closed won new acquisition opportunities or have significantly altered your product line, you may need to look at a different data set. In these cases, it may make sense to focus on qualified opportunities created in the last few months to increase your minimum data set.
If you’re brand new or opening a new market, take your best guess at your ideal customer profile and form a group of target accounts as a starting point for your model.
Once you’ve identified your “ideal” population, pull data related to these records and determine whether or not there are patterns. Look at firmographic (size, revenue, industry, subindustry, region, etc.) and demographic (title, department, role level, tenure, etc.) data values.
If you’re a larger organization with different product lines (perhaps you sell both B2B and B2C packages), developing a single ideal customer profile wouldn’t make sense. Your data would be muddled with such a wide range of buyer personas. At a minimum, build a separate model for B2B and a separate model for B2C purchases.
It may also be beneficial to layer in models specific to expansion and renewal sales if your company is emphasizing customer retention over net new acquisition (or perhaps they want to do both). Regions can also throw enough variation into the mix that splitting the model by area may make sense.
Define and collect firmographic and technographic about your customers and prospects. Don’t neglect harvesting niche signals in your customer market. For example, if your product works best with certain operating systems or if your product appends to a specific technology, make sure to enrich your data with a third-party data source that provides this critical information.
02 Analyze Your Data Points
Once you’ve identified your ideal audience, determine whether your data is collected during the course of the sale. Let’s say your Use Case field on the opportunity is required at 40% probability and isn’t filled out before then. Including Use Case in your ideal customer profile scoring model wouldn’t make sense.
Look at whether or not the data is populated for a significant portion of your database. If your executive team knows that IT Managers with a team of more than 20 employees are more likely to buy, but the team size is only populated on 10% of your records, it may not be meaningful. Look at whether the total population of accounts is likely to convert at a lower rate than those with the value filled out. If the difference isn’t significant, don’t use it in your engagement model until you change data enrichment tactics.
Run a regression analysis on each of your variables and compare them to your customer base. If there isn’t a strong correlation to conversion, don’t use the variable. When comparing multiple variables to your customer base, you want to see that more matching variables increase your chance of conversion versus fewer matching variables.
03 Building Your Model
Usually, you have a few options when deciding where to house your ICP model. Most marketing automation platforms allow for demographic scoring. You may need to use some creative mapping to get the values you need for your model. Your CRM is always an option, and you may use a data platform and push calculated scores back into your MAP or CRM. Just be certain that the score is available in the systems your sales and marketing teams will be using.
If you add up each of your data points for a total potential score of 100, do another analysis to determine how quickly your conversions fall off. You may only have revenue associated with the top 25% of scores, and that will change how you split your model into grades. In this example, A may be scores of 90-100, B may be 83-89, C may be 75-82, and D may be 69 and lower.
Generally, we recommend that people implement their scoring model and use it passively while it’s being tested out. We’re a fan of validating things before potentially gating leads. For example, you could use the ICP score to create views in Salesforce or Outreach and get sales feedback on how prospecting goes in each of the tiers.
If you use a machine learning model and have a reasonable volume of data, try limiting the time range of data analyzed to the last one or two quarters rather than the lifetime of the product. If products are redesigned, it may impact your scores. If the market significantly changes (which it certainly has during the COVID-19 pandemic), looking at older data sets won’t be of much help.
04 Provide a Feedback Mechanism
There will be major market changes that can negatively impact a scoring model. Give your sales team a feedback mechanism and review the feedback on a monthly basis. If you begin to receive a higher volume of negative feedback, it’s time to do some investigating. Did the product change? Did something happen in the market? Your model may need to be adjusted to match your new circumstances.
05 Schedule Quarterly Reviews
A feedback mechanism is a great way to gauge how people feel about your model. Analyzing the data to determine whether or not your model is still statistically accurate is vital.
Your analysis should answer the following questions:
- As are still converting better than Bs?
- Are Bs still converting better than Cs?
- Are Cs still converting better than Ds?
- Did this quarter’s converted leads and opportunities produce pipeline and revenue as expected across your ICP?
If any of these factors changed, you might have to rerun your analysis and make adjustments.
Calculating your ideal customer profile is more complicated than people would imagine, but it makes a massive difference across both marketing and sales efforts if it’s done correctly. Your campaigns will be more efficient, and your sales team will need fewer at-bats to produce a larger amount of revenue.
Let us know if you have any questions or comments!