Try a creative approach to ideal customer profile refinement

By Jason Widup , VP of Marketing at Metadata

When I started working at Metadata three years ago, we were still a startup figuring things out. I guess we still are in a lot of ways! One of the things we couldn’t figure out was how to tell the difference between our best and worst customers. The problem was our broad ideal customer profile. We can sell to any B2B company that advertises on LinkedIn.?

In a basic attempt to tighten it down, we looked at who bought from us the fastest. The answer was 150 - 2,000 employee B2B software companies, so we started selling to as many of them as possible. If you were a software company in this employee band, you were in our ICP. Still thousands of companies, not necessarily a target list.

A year later, we saw churn and dug in to figure out why. Using common third-party firmographic metrics like company size, industry, headquarter location, and revenue bands, all of the companies looked the same, whether they churned or not

We took two creative approaches to refine our criteria and improve the experience of discovering Metadata for our target accounts.

Finding a new data source

Knowing that available third-party data gave us only a high-level view of our ICP, we turned to internal data with the goal of finding something in common between our best customers. We hypothesized that the amount of money customers spend on advertising on certain channels through our platform may be a differentiator.?

Sure enough, customers spending more on those channels stayed on longer than customers spending less. From there, we identified a monthly spend number as a cutoff. Below that number, on that one factor alone, we considered customers churn risks.?

So now we had done step one: We had a clear behavior that was present in our best customers and absent in our worst. We could use that cutoff to help find more prospects in our ICP. We just needed a source of data to get their ad spend on those channels, particularly LinkedIn, before they became real prospects.?

Finding a source quickly became an obstacle. There are third-party resources for Google ad spend, companies that use proxies, but nothing for LinkedIn because LinkedIn is a walled garden. Nobody can sniff its ads the same way you can Google ads on websites or search engines.?

I was able to find a solution by connecting a few dots when we ran a promotion where we were offering a free audit of LinkedIn ads for prospective customers. I realized you can click from a company’s posts to very easily see the ads they have loaded up in their account.

Now, I wondered if there was a proxy between the number of ads I could see and their spend. If I could find a correlation to spend, we could use it to prioritize target accounts that spend like our best customers.?

LinkedIn shows a mix of live and expired ads, and it’s not immediately obvious how they chose. I’m still not entirely sure, but I was able to commision a company to scrape the count of ads for a list of our customers. Knowing what our customers spend, we wanted to see if there was a correlation between count and spend that we could use as a proxy for spend in prospects.

The count and spend give us a proxy. It wasn’t as simple as something like five ads means an account spends this exact amount, but we could create buckets. We found that eight ads was the minimum number to qualify as ICP.?

We had to repeat the process to scale it as an ongoing resource. We would refresh every three months, loading the data into Salesforce for the sales team to see. The change over time was recorded as well, so we could see if companies were spending more.?

Try it!?

You’re probably not looking for LinkedIn ad spend to add to your ICP criteria, but you can still use this process.?

1. Identify the uncommon differentiator(s) between your best and worst customers
2. Find a source of data that speaks to the differentiator so you can apply it to prospective customers
3. Decide how to scale the process

Looking at an existing data source in a new way

Although third-party data didn’t tell us a lot when we looked at it as a list of attributes, we found a way to use it to refine our ICP as well. Sometimes you just have to look at readily available data in different ways.?

Marketing to companies that have recently raised funds is a popular tactic because fundraising is commonly understood to signal money to spend. Rather than simply look at companies who are crossing promising milestones today, we took the idea a step further, reverse engineering from our existing customers to decide which signals were most likely to be meaningful for us.

We looked at the news cycles coming out of our best customers three months before they became our customers, choosing the timeframe because our sales cycle lasts three months. When they started the process, did they have recent funding? What else was going on with them??

We were able to essentially recreate intent data that signaled strong conditions to look for in prospect companies. Aside from funding, we identified free trials, G2 badges, and open demand gen roles as indicators of high potential.?

Try it!?

Again, you can try this process even though the specifics will look different for your business.?

1. Identify the signals your customers sent out before becoming your customers, using the length of your sales cycle as the amount of time to go back
2. Find a source of data to sort prospective customers according to the same signal
3. Decide how to scale the process

What we did with the new ICP matches

Beyond prioritizing accounts that signaled the most likeness to our best customers, refining our ICP allowed us to increase engagement with our marketing experiences and to better score accounts.

Personalization

Now that we knew how to look at a target company’s LinkedIn ads, we would compare their output on LinkedIn and Facebook. Often, they had 80 to 90 percent of their ads on LinkedIn and 10% or less on Facebook.?

We personalized outreach and ads with the exact number, which really catches someone’s eye. Then we would tell them we know they don’t spend much on Facebook because B2B targeting on the platform isn’t optimal, and explain that Metadata unlocks better targeting.?

Whenever you market with personalized information, there’s a fine line between being impressive and being creepy. It’s possible to find home addresses, but reaching out that way is probably going a little too far for? a software company. Trust your gut.

Removing qualification obstacles

To optimize our revenue, we only send demo requests to our sales teams, not content leads. We grew off of incentivized demos and I knew we didn’t want to pay somebody to get a demo if they really didn’t meet the qualifications of our ICP.?

For that reason, our forms would qualify people on ad spend and whether they use Salesforce (because it’s the only CRM we integrate with to optimize against revenue).

Once we got comfortable enough with our proxy for ad spend, we could skip that qualifying question, removing an obstacle. We were confident enough in the data to trust it and we’d pad the low end of our proxy to make it more reliable.?

Since we did this work, we’ve developed a more sophisticated account scoring algorithm that this work fed into.?

Get your ideas flowing

The creative ways you reverse engineer new ICP criteria will be unique to your business. To get you started, I brainstormed starting places for five different situations. Take them as inspiration and generate a list to try.?

Healthtech - AI for medical coding that makes it faster to enter and read chart information

Healthtech can be challenging for privacy reasons and also you’re often selling to completely different people. You want to sell this to smaller private practices. There is a government list of all private healthcare providers in the US called MedlinePlus that can be scraped. Tap into insurer lists of healthcare providers.

Cybersecurity - Cloud-based email protection that detects targeted attacks and phishing attempts before they compromise a business?

• CISO’s and other privacy and security professionals make it a habit to be hard to find! It’s like a game to them. Most of them have burner phones and private VPNs :).?
• These folks tend to hang out in private forums and groups, and they read journals in their field (online and off). You can advertise directly into some of these and others you can get lists from.
• Ask your company’s CISO where they hang out and how they find out about new technology in their field. These tend to be event-heavy as well, so sponsoring an event where you get the full list could work.

Fintech - B2B invoicing and payment app with customizable payment requests and processing?

• Fintech is also heavily regulated and the buyers are often not consistent from company to company. Sometimes fintech companies have their own restrictions on how you can collect and use data, even internally.
• This one can be more straightforward because these buyers tend to be on social networks and there tends to be a lot of paid search volume for the problems they solve.
• For an interesting look at prospect accounts, leverage finance job opening info, recent funding info and other events that could spell the need to manage finances better.

Augmented reality - Smart cameras and machine learning that improve B2B communication and collaboration?

This is one where you could get by with scraping case studies off of websites of competitors or complimentary products. For example, Microsoft is making a huge push into augmented reality with hardware, software, etc. They have a lot of case studies across their websites. Understanding which companies are really adopting Microsoft’s tech and then creating lookalike accounts from those could be a great place to start.

HR tech - Gives startups and small companies access to the HR, payroll, and benefit tools big companies rely on?

• Again on this one, job opening/hiring data can be helpful. Looking at the change in number of open positions could give you a sense of which companies are ramping up hiring and may need a new HR tech solution.
• Series A and B funding rounds are also often when companies have to move from a very basic and immature HR product to something more mature. Leveraging funding data and timing can help you here.

Let me know how it goes! I’m easy to connect with on LinkedIn.