Improving our freemium by better understanding the cost of support

Freemium models usually have a bad reputation within SaaS companies. You don’t want a user not willing to invest a few dozen dollars on your service. Or do you? At Email Hunter, we found that having a generous free plan brings several advantage to our business. I won’t be focusing on those today, but rather on what we see as the biggest trade-off of this model: the amount of support we get. If you need 90 free users for one paying customer, that’s a lot of additional emails every month!

The cost of customer success is very real. For us, we have dedicated one member of our team. But everyone also spends some time answering questions on a daily basis or assisting on technical issues linked to their work. Thousands of dollars get spent each month on support. There’s no doubt it is worth it, though. Great customer success means higher conversion rates and helps for upsells. Even users that don’t end up paying can help by reporting bugs.

As we continue to grow, we want to handle support in a more conscious fashion. Our goal is to handle more users while keeping the size of our team small and manageable; this means changing how users get support from us.

Potential solution: transferring the cost of support to users

Support isn’t free. Why not only make it available to paying users? The standard approach is to either only help free users with bugs (but that’s the best way to frustrate many people), or change the response time depending on the user. This last approach isn’t feasible for our size and also leads to frustration.

Making users pay for support also hurts conversion rates. Many would-be customers have a small question that needs an answer before they can upgrade. We obviously don’t want to make it tough for them when they’re likely to convert.

Adapt support depending on the country

We noticed we get more support inquiries from some countries more than others. Without looking at the data, it seemed some didn’t convert well. To be sure, I took a dive into our Analytics database. We’re using Segment to save events, and store everything within a Redshift database; a great thing Segment does is send Intercom and Stripe data. It makes it possible for us to cross financial and support data. To easily generate nice graphs 📈, we then use Mode Analytics.

This graph displays the percentage of both MRR (Monthly Recurring Revenues) and conversations received by country. Our goal was to get a list of the countries requiring significant support resource but not generating any revenue. In this case, India, Brazil, Singapore, and the Netherlands cost us a lot more time than they generate revenues.

What can we now do with this list? As I’ve said, we’ll always provide support, but we can change the way users get access to it. For now, chat is available at the bottom right of every page. For some countries, we can disable it and force them to use emails. We can also make sure they always go through our Help Center before contacting us. We’ll be testing various approaches shortly. We’ll share the result of those experimentations later in the blog!

Identifying users most likely to convert

Some users are more apt to convert depending on how they use your service; this is the case for our API users. Therefore, we’ll make sure to continue providing them with very easy access to support to boost the conversion. It should make support a lot more efficient and profitable.

Users of a particular part of your service that aren’t converting might behave this way because the product isn’t good enough. Disabling support might take some of the disappointed messages away, but it shouldn’t stop you from fixing what needs to be fixed. Because of this, we don’t plan on stopping support for users depending on how they use our service.

Conclusion

We know using data will help our company continue growing while maintaining a successful freemium plan. Of course, there are other ways to optimize customer success, and I would love to hear yours! Feel free to share your thoughts on Twitter at @antoinefink or email me: antoine@emailhunter.co.

P.S. If you’re using Stripe, Intercom, and Segment’s sources, you get your graph in a few seconds. Here is the SQL code that I used. It can be profoundly improved! Don’t hesitate to send me your upgraded version and I’ll update this one.