Optimizing self-service SaaS models... with sales reps

Self-service. Oh, show me a SaaS founder who doesn’t love self-service customers. They find your product through word of mouth, inbound marketing or partnerships. They eagerly sign up for your trial. Automatically convert into paying customers. And no member of your team needs to sell or support them.

It rarely happens this way, especially in B2B - but even if it does, you still should employ sales people. Even if your product is "so good that it sells itself", you should have a sales team that actively sells.

"We have no sales people." (Except those we don't tell you about)

There a very successful SaaS company that proudly proclaims they don’t have sales people. Let’s call them NoSalesCorp. I admire this company, I think their strategy is brilliant and I love that they're so product-focused.

But the truth is that they do have a sales team! While they’re not using sales to directly grow revenue, they do use their sales team to establish benchmarks against which they measure their self-service funnel.


A/B testing self-service vs sales people

Most of their users and customers sign up for their product without ever speaking to a sales person.

But their "secret sales force" reaches out to selected verticals and cohorts of their users.

NoSalesCorp then measures how well their self-signups convert in comparison to users who’ve been contacted by a sales person, what the lifetime value of those customers is, how successful they are using the product, etc.

And when the sales team cohort performs better than the self-signups?

They call it a bug!

A bug in their "product" (the onboarding/marketing/UI) that needs to be fixed. They then deploy their product team to come up with ways to fix this bug.

The product team analyzes what the sales people exactly did that made the difference.

Then they run a series of experiments and measure how every change affects the performance of the self-service process, benchmarking it against the sales people’s performance.

Running experiments to optimize the self-service process

There’s a nearly infinite number of things they can do to "fix the bug", among them:

  • Adding disclaimers to alleviate concerns at certain steps? (e.g. assuring that they won’t send tweets from your account if they connect their twitter account, or that they won’t pull data from your gmail account without explicit permission).
  • Adding/more prominently featuring/removing the option to skip a step in a process.
  • Do they need to tweak something in the user interface, e.g. adding a process bar?
  • Do they have to change the wording of a drip email? A headline? A subhead?
  • Can visual elements be changed that’ll convey the value in a better way?
  • How can they better educate and sell users on the product?
  • Where can they remove friction from the activation flow?
  • Do they need to add a step in the onboarding process?
  • Adding/removing/changing push notifications.
  • Do they redesign the documentation?
  • Improve their call to actions?
  • Adding social proof elements
  • Removing form fields
  • etc.

This isn’t even beginning to scratch the surface of what’s possible. By employing a sales team and keeping track of their interactions with customers, they constantly get new data points on things they can do to improve their self-service process.


Having an internal sales team can help you build a better self-service sales funnel. Simply measuring the performance of your self-service signups versus those that signed up through a high-touch sales process will provide you with valuable insights on how to optimize various parts of your self-service model.

Table of Contents
Share this article