Understanding customer behavior and intent has become a crucial competence for pretty much any business and the only way to gather that knowledge efficiently is to automate as much of the process as possible.
But that introduces a risk. Poorly executed automation can be alienating. In an era where companies want to build lifetime relationships with customers but those same customers are increasingly fickle, is the trade-off worth it?
The good news is smart technology and process choices can enable you to gather the data you need without making extra work for your customers.
In this installment of our series on automated customer communication, we’ll look at the benefits and pitfalls of automated customer surveys, followed by a glimpse at what the future might hold.
Every Day’s a Survey Day
How companies perform customer research hasn’t changed much since the web became a significant part of everyday life. Most of it takes place in one of four ways:
- Web-based: often the user is interrupted and invited to take a survey
- Voice calls: either human-powered or increasingly automated
- Messaging: an automated SMS conversation asks questions, usually one at a time
- Email: usually the participant is invited to visit a web-based survey.
However, the frequency with which companies ask for feedback has changed. As a consumer, you’ve no doubt noticed that companies are suddenly very interested in what you think of them.
Twenty years ago most market research took place either with clipboards on busy shopping streets or with panels of people who had been selected to create a group representative of broader society.
Now, almost every purchase or interaction with an organization triggers a “How are we doing?” email. As someone working in customer communication, you no doubt understand why things have changed.
We’ve gone from a largely transactional economy to one where companies want to build long-term relationships with customers.
If you ran a hardware store in 1998 and you sold a pack of screws to a new customer, you’d have little to no idea how much it cost you to acquire that sale and you probably wouldn’t know if you’d see that customer again.
Today, things are different. Here’s why:
- Customer acquisition costs are now transparent.
- Competitive ad markets have raised the cost of acquisition, meaning that it takes more than one transaction to turn a profit from a customer.
- We capture customer contact details as a matter of course.
- We have the analytical tooling to show us how even small tweaks affect turnover and profitability.
If a new customer buys a pack of screws from our online hardware store, we can:
- see precisely how many more sales we need to pay-off the customer acquisition cost.
- tune the customer experience to increase revenue per customer.
- contact the customer effortlessly.
Like a puppy craving attention, the temptation can be to keep asking the customer, “Do you love me? Am I a good dog?” over and over. Newspaper comment pieces suggest that, at least amongst journalists, some resistance is appearing.
Pitfalls of Automated Customer Surveys
So, we have a dilemma: we want to know how best to serve customers and we believe that they can tell us something about that. Customer surveys are low-cost, easy to manage, and gather enough responses to draw comparative conclusions.
On the other hand, our survey is likely to be one of many that a customer sees that week. This means our survey is likely to be buried and ignored or, worse, deleted in annoyance.
In fact, there are some downsides to automated surveys beyond survey fatigue:
- Self-selecting audiences tell us about the people most likely to answer a survey and not about the customer base as a whole.
- Asking the right questions can be hard.
- Not everyone will answer honestly or even be aware of how their biases influence their answers.
If we ended the story there this would be a pretty downbeat blog post. Thankfully, you can learn about your customers without annoying them.
So, what does it take to do automated customer surveys well?
Back to Basics: What Will Automated Customer Survey Responses Tell You?
The survey is the hammer of the customer insight world. Think about home repairs. You can deal with most home repair problems using a hammer but a hammer won’t necessarily solve those problems; you might just give yourself a feeling of dealing with it while you’re actually making it worse.
It’s the same with surveys. If you have a question, you can throw a survey at it and you’ll get an answer–but will it be the right answer?
First up, you need to be clear on what you are measuring. Listing the many types of survey you could run would take a blog post of its own, but broadly your motivation for running a survey is likely to be:
- understanding who your customers are and what they want
- testing your current product or validating a future product
- measuring customer satisfaction.
You could research all of these with surveys, but perhaps another method would be more effective. Or, maybe, there’s a different way to run a survey.
Selecting Participants for Your Automated Customer Survey
Like all data, surveys suffer from the GIGO problem: garbage in, garbage out. One factor that leads to “garbage” data is selecting the wrong group of people to take part.
Full-time market researchers have sophisticated ways to select precisely the right survey population. They’re looking to make sure the sample size is large enough to be statistically significant, and the participants are representative of the larger population being investigated.
Automated customer surveys can feel a little crude in comparison.
Displaying a survey pop-up to every tenth website visitor, for example, won’t randomize your participants. Instead, the people most likely to respond are those who are unusually unhappy or bored, thereby skewing the results. And, seriously, how happy is anyone going to be when their visit to your website is interrupted by a survey pop-up before they’ve even been able to do anything useful?
But today you know more about your customers than ever before and that means not only do you know what a representative audience looks like but you also have the data that allows you to automatically select representative respondents. You can even tune who you target as you go, according to the types of customer who remain under-represented in your responses.
So, once you know who you want to hear from, how do you start the conversation?
Inviting Survey Participants
As an industry, we’ve become adept at creating experiences to delight customers. And then we stomp all over it with a survey invitation pop-up or yet another email.
We can do better.
Contextual communication has revolutionised the effectiveness of customer communication and it points the way here. As Roland wrote, contextual communication draws on what the customer is doing, where they are, and what we’ve been able to derive about them from the data we hold on them.
For example, if your customer has been using your mobile app and appears to have completed their task, an SMS inviting them to answer a question takes advantage of your knowledge of what they’re doing (using your app) and their physical context (using their mobile). Or perhaps you have permission to communicate with the customer using WhatsApp, meaning you could send your survey question using Nexmo’s Messages API.
The key here is to approach inviting your survey participants using the same skill and nuance as your other customer communication.
Conducting the Survey
If contextual communication offers a better way to invite participants, then it also shows a better way to conduct the surveys themselves. Web-based surveys have their place, but you can determine which customers will respond best to them based on your understanding of how they interact with your company generally.
For other customers, you have to find the right channel:
- SMS, WhatsApp and other messaging services
- automated phone calls using text-to-speech to ask the questions and speech recognition to capture and respond to the answers
- human-driven calls for more involved surveying or for those customers who respond better to human interaction
- in-app messaging
If you see customer surveys as another part of your contextual, omnichannel communication strategy, then it becomes clear that you must meet customers where they are.
What Comes Next for Customer Survey Automation?
As technology continues to open new ways of understanding and communicating with customers, how can it help to evolve the humble old survey?
The dominant thread of any conversation about the future of customer communication inevitably comes back to AI. In particular, AI-powered virtual assistants seem set to augment human-led customer service. Imagine a survey that becomes a spoken conversation with a friendly assistant, rather than a series of dry questions.
Or perhaps AI and greater analytical capabilities mean that the future of the survey is limited. Management consultancy Bain, for example, suggest that better analysis of customer behavior may be the future of customer feedback; so in the future we might put more effort into actively observing customer behavior rather than asking them questions.
Whatever the future of automated customer surveys, today we need to meet customers where they are in order to gather the most useful survey responses. When it comes to surveying customers, contextual communication will help you to find the right respondents and engage with them in a way that meets their needs.