How we used surveys and emails to get to know our audience

Lorenzo is the marketing & community manager at pr.co. His work is to help people get the most out of their PR efforts online. He also uses aggregated data to improve user experience on pr.co and to find patterns and trends in the PR world.

In March 2014, we developed a strategy to understand why the customers of our SaaS startup chose us and where they saw the most value in our toolkit. We set up a survey and emailed our mailing list to ask our users to take it.

This is what we did, step by step, and what we learned in the process.

A little introduction: pr.co is a SaaS startup that allows you to create, publish and distribute your press releases online. You can customize your pressroom, add your presskit, and get real-time results when you send emails to your media contacts.

Pr.co was officially launched in November 2013 and it’s made of 3 people, Dennis, Jeroen, and me. While we use to talk to customers everyday, we recently, got busy developing stuff, adding features and rebranding. At some point, we realized we wanted to check in with our customers to see whether we still were on the same page.

In the last 3 years, we were able to build an interesting mailing list of 10,000 people. We use Mailchimp to keep in touch with them, about twice a month. We regularly experiment with our newsletters: sometimes we just explain new features or make special offers, other times we propose interesting content we write or read, and we always do A/B testing on the emails’ subjects.

In March, we asked our mailing list to take our survey. This is what we wrote:

This is the usual tone of voice we have in almost every email. The survey from Google Docs was also very simple.

As we wanted to gather some qualitative information about our users, we mixed multiple-choice with open questions. Thus, questions about what they liked were mostly multiple-choice, while those regarding dislikes were open as we wanted to know what they didn’t like in their own words.

From multiple-choice questions we were able to gather data and create charts about the value the users get from different features. From open questions we learned more about the way our users speak and what words they’re using when referring to us.

Since we were about to release a new design for the press releases, we also wanted to understand how many of our users had this need:

The results of the newsletter were okay. The open rate was quite good compared to previous blasts, but the click rate sucked. But it’s just because everybody hates surveys, right?

Eventually, we gathered around 40 responses to our survey and the insights we got from them were priceless. We learnt that almost 50% of our users are part of a startup (or founded one), but startups are not our first source of income (yet).

One of our most internally debated topics is the pricing model. We aim at committed users that want to invest some of their time to get the most out of our toolkit. So we always try to show the benefits of a yearly subscription, which is also 20% cheaper than its monthly equivalent. That’s why we asked how often our users were using pr.co: the price per release of a twice-per-year user is considerably higher than the price per release of a twice-per-month user.

When we emailed our users to ask them to take the survey, we said “You’ll get a better product and the results will be shared” and we meant it. So we wrote a blog post about the aggregated results we got from the survey and shared it in another of our newsletters.

We also replied directly to some of the users who took the survey, when we felt the need to ask more. In the survey we asked if it was ok to get back via email, so we were not afraid about it. If some users did not know about a specific feature we should have told them about, why not letting them know with a simple email?

Right after we got the results of the survey we gathered all the feedback we received and we split the tasks into different actions: some issues could be addressed right away, some required a little more time, some others were not in our schedule.

So we replied via email to some users when we felt the need to let them know something they may have missed; we set up a list of nice-to-have features we will be working on in the next few weeks; we also set up an excel sheet with all the complains we get, for further reference. Finally, we started a new experiment involving a fake feature to see how many people would like to have it. But this is another story.

How do you get to know YOUR audience? What can we do in a different way next time?