4 Steps to Knowing What Your Customers Want Better Than They Do
The problem with asking customers what they want, like, or think is often that:
- They don’t know.
- They think they know, but they’re wrong.
- They know, but they don’t tell you the truth.
Companies famous for predicting customer behavior (like Amazon, Netflix and Pandora) have discovered that “implicit data” (e.g., observed customer behaviors) is much more reliable than “explicit data” (e.g., customer-provided information). Indeed, to predict customer behavior, the most sophisticated companies prefer to observe customers rather than listen to them.
How do entrepreneurs get in on this action? Fortunately, business owners now have an array of data and tools to help predict customer behavior, even without the budget and resources of big companies.
Here are a few ways to utilize quality data and know your customers like never before:
1. Reduce surveys and “be honest with me” conversations. Surveys take a ton of work. You beg people for their time, an unrepresentative group responds (i.e., the ones who like you). You don’t get enough responses to attain statistical significance and results often prove that customer intent doesn’t match future behavior.
Direct conversations with friends, advisors and test users are usually misleading as well, because not everyone will be honest. It’s human nature to avoid disappointing an enthusiastic entrepreneur about the product that represents their blood, sweat and tears. Often, you end up with feedback and data that’s falsely optimistic and positive.
2. Get accurate, unfiltered customer feedback. Support and observed user testing are a couple great sources for useful customer feedback. Customer support calls and emails are goldmines. People give you direct, clear and candid feedback. This is the blueprint for their upcoming behavior, often revealing things they don’t even realize.
A painful but effective feedback source is a service like Usertesting.com, where you can observe someone using your website or mobile app, viewing your product or watching your demo, as they narrate their thoughts in real-time.
This unfiltered feedback is invaluable. These people don’t pull punches, so while the comments may seem a steady stream of downers, don’t resist this tactic in favor of asking a friend to give you “honest feedback.” One way helps you build a successful business. The other only helps you feel good about yourself.
3. Observe real actions. What people do is always a better predictor than what they say, so analyze their actual usage on your site. Google Analytics is a great starting point, but you should move beyond it to understand individual and event level usage. Tools like Mixpanel or KISSmetrics, give you easy-to-understand data and visuals that are gold for our next step, which is predictive analytics.
4. Take a little step into big data. With customer support, user testing and better site usage data, you can start to replicate the efforts of Netflix, Amazon and Pandora to predict customer behaviors. By running all the text from customer support calls, emails, chats and recorded scripts from user testing, natural language processing tools can analyze what words and phrases really signify.
You may find that, “This is the last time I,” means a customer can be saved, “sick of“ means they need a special offer, and “This is the last time you,” means they’re leaving. You can also track what language means in the longer-term (e.g., “hate” may mean the customer will stay, but only for a couple months).
Machine learning tools will also process site usage data, analyzing current behaviors to predict future behaviors. You can combine multiple customer data sources to see more correlations. People who use a certain credit card may be infrequent customers. People who upload pictures and are under 30 may refer friends. People who subscribe to your newsletter and say “I love your…” in customer support emails may buy your highest priced products.
Every entrepreneur is taught the old adage, “The customer is always right.” But analytical tools and predictive analytics are proving they’re sometimes wrong. Stop relying on what customers say to shape your business and instead focus on the data they generate -- finally unlocking the power to know them better than they even know themselves.