Calculate the true cost of supporting your free users

As the creator of a self-service channel and driver of growth for a B2B SaaS tech startup, I work closely with support agents to optimize their processes and make their lives easier. Good support, self-service and of course, user experience, are all part of improving customer satisfaction and retention.

Although these free users can eventually expand, depending on your product offerings at the free plan vs paid plan, many of them may never move to paid. This means that you have a higher Customer Acquisition Cost (CAC) and a lower Lifetime Value (LTV) — CAC is higher with free users on board, while LTV is essentially zero for the entire span of days that a free user remains non-paying. Those customers are actually costing you money!

To get to the bottom of this, I enlisted my trusty confidante and work spouse, the head of support. I’d started bonding with support when building the self-service community to add to the Help center I created, thus I’d read literally hundreds of support tickets and worked with two awesome agents who helped me get content ready so that I could seed a new community every day for a year. I was intimately familiar with the pain that customers were encountering — knowledge that helped me to create a better product, as well as to optimize support and self-service processes. This was the next step in a number of improvements in building up self-service.

How many users are free users?

After determining the cost of the tickets — we’ll change the number here to $50 per ticket — we then had to determine how many of these were free users. While this might seem relatively easy, since this was a startup, they had never built for data analytics, so there were many data gaps and lack of integrated data to make decisions.

Data challenges:

Here were some challenges with the data:

• Although I could tell when an account’s free trial had ended, this information was not sent to the support application.
• There was no free trial start date being captured at all.
• The licensing model had changed a few years into my tenure, thus there was no data consistency, as a user in the old model did not have a platform subscription, thus could appear as if they were free. There was no easy way to determine how many free users there actually were.
• There could be any number of users on a free account.
• An account was rarely marked as inactive, so it was difficult to know which users were active users that should be counted.
• There was a lack of data integration, so even the sparse data that existed in other places was not available to the support UI, which added another layer of difficulty in tracking down the cost of free users.

Never one to shirk a challenge, I began at the easy end of things — with the new licensing model, let’s say there were 50K free users. This is after removing users that are still on a free trial. In the old licensing model, we’ll say that there are around 18K users. We’ll ignore those for now, and concentrate on the low-hanging fruit.

How many free users are entering support tickets?

For these 50K users, the next thing to determine was how many of them were entering support tickets. This hasn’t been implemented yet, but here is the plan.

1. Compare the list of free users against the list of users opening support tickets. If there’s been a ticket opened in the past few months, consider that user an active user.
2. Once we have the list, sum up the tickets being submitted by free users, as well as the average number of tickets per free user. In this case, we’ll say that the ticket density for all users is 0.5 tickets/week. That means that each user is submitting at least 1 ticket every 2 weeks.
3. For the purposes of this hypothetical value set, let’s say that 40K of these users still appear to be active (by looking at Monthly Active Users (MAU) in an authenticated analytics tool where you can get actual user details). If each ticket costs the company $50, and even half of those users are entering 1 support ticket every two weeks (so 24 over the course of a year), doing the math, we get 50 * 20,000 * 24 = $24M spent on supporting these free users, or $1200 per free user per year! This is a massive number and one we definitely need to pay attention to!
4. From this analysis, the next step is to make sure that your self-service channel is well-provisioned with content and with human experts so that you can try to shift this cost from $1200 to almost zero (I say “almost” because self-service includes community posts for help, which takes up expert time to respond, but not all posts are responded to internally, if you’ve built a solid, engaged community). By moving free users to self-service, you get the benefit of freeing agents up to keep paying users happy, improve your understanding of customers (as many more will be posting on your community rather than going to support, which builds awareness and drives prioritization of usability issues and functionality gaps), and giving your free users an incentive to move to paid, thus deepening their commitment to your product and company, as well as their determination to extract the most value out of your service as possible.

To make this dollar figure more accurate, you could also determine the number of tickets submitted by paid users who then converted to free users, or vice versa. Subtract the number that they submitted while paid from their total before you average the total number of tickets per free user.

I recently completed a certification in growth management, which has opened my eyes to a world of possibilities on quantifying and reducing costs, as well as acquiring, activating and retaining the right customers, so am sharing how I’ve applied my learnings. If you have other ideas on areas that you’d like me to write about, let me know!