September 24, 2015
Automation Experiment: Solving Customer Problems With Proactive Emails
At Autopilot we’re focused on making easy and visual marketing software for, you guessed it, marketers. But we’re also big believers in “eating our own dog food” and using our own software to discover new ways to improve the customer journey. Not only does this ensure we’re using Autopilot as much as possible, which helps us to refine the user experience, but it’s also really fun. In a series of posts I’ll be showing you some examples of experiments we’ve executed using Autopilot to help spark ideas for your own customer journeys. The first of these is an experiment based on a question many companies have no doubt asked themselves at one point or another…
“Are people getting the answers they need from our knowledge base?”
As you know from when you’ve searched knowledge bases yourself, if you don’t find the answer, often you’ll just give up. Sure, if you’re really proactive you might log a ticket and ask for help; but not everyone is willing to put in this level of effort. As a result, you might be missing your chance to help new leads, or to ensure that existing customers can easily solve their problems, both of which grow customer lifetime value. We use Zendesk, which gives us a report of commonly searched terms. This is helpful, but it doesn’t necessarily give you a deep insight into what problem the person searching was trying to solve. It’s also not personalized, so you can’t know who needs help.
The personalized email behind the experiment
So our experiment was to set up a smart segment in Autopilot for anyone who visits the search page on Zendesk. When they enter this segment, we notify ourselves on Slack for maximum transparency. We then wait 30 minutes, and send a personalized email asking if they found what they were looking for. The 30 minute delay is to avoid emailing someone too quickly; it wouldn’t be as effective if the person received the email the moment they performed their search. We also configured the email so it wouldn’t be sent again if they came back another day and performed another search; we didn’t want to be too pushy, after all.
The gamble: was this creepy or helpful?
When we kicked off this experiment we thought it could go two ways: either people would think it’s creepy, or they’d appreciate the help. So we took the gamble and luckily the latter was the case. In fact, the majority of people went out of their way to express their appreciation of the offer for help. In less than two weeks, we sent an email to 96 people who searched Zendesk. Of those, 61 opened it (63.5% open rate) and 25 replied (26% reply rate). Two were existing customers and the rest were people who are on free trials. Of those 25 responses:
14 couldn’t find the information they needed to answer their question
2 said they found what they were looking for
5 wanted to know if Autopilot had a particular feature
4 wanted to setup a demo call to learn more about Autopilot
The feedback we gathered from this process was invaluable, and resulted in us adding several new articles to our knowledge base, and making several others more clear. It also enabled us to start conversations with leads and gather product feedback that might otherwise have been missed.
Try this experiment for yourself
So there you have it. I strongly recommend you give this a try if you have a knowledge base or a frequently asked questions page. It’s an experiment that has opened our eyes a great deal, and is one that we’re continuing to run. The benefit of using automation for an experiment like this is that once it’s setup, you don’t need to do anything until you get a response from a customer. In fact, we’ve made it even easier for you to try this experiment by adding it to the Autopilot Guide Book. Good luck, and stay tuned for our next post in this series.