Adding Friction to Drive More Sales

Best practices are clear. In a purchase funnel for an e-commerce app, reduce friction as much as possible. Learn from Amazon’s one-click buy ! Reduce the time the user spends on your site before they put in their credit card. Every click your site requires is an opportunity to lose a customer.?

This common, broadly accepted truth helped improve the user experience for a myriad of brands. Guest Checkout became commonplace thanks to it. Ability to check out using Apple Pay. Auto-complete of addresses. Each of these is a tiny pain remover; customers, with their busy lives, are more likely to complete a shorter task - instead of being distracted in the middle of it.?

But while many friction reducers are a net positive for both the customer and the e-commerce brand, is it fair to claim the converse - that any addition of friction is, relatedly, a bad thing for conversion? Not so fast! This depends on what you’re selling, and where the customer is in their consideration funnel.?

At Hungryroot, the company where I work, we’ve seen a fascinating phenomenon. As a reminder, Hungryroot is a personalized grocery service: tell us about your family’s dietary preferences, and we’ll pre-fill your cart with groceries and recipes you’ll like, and deliver them to your door.?

Our signup experience includes a hefty number of steps where we ask the customer about their food likes and dislikes. The common wisdom goes: reduce the number of these steps as much as possible - and definitely don’t add any - to maximize conversion.?

We decided to try out a contrarian hypothesis. We launched an experiment where we added a question to the signup funnel: we asked about the customers’ favorite proteins. Lo and behold: conversion increased. It’s as if this new funnel step had an individual conversion rate of over 100%!.. How can this be? We dug in and discovered these key reasons behind this counter-intuitive outcome:

1. We are selling a non-commodity product. The exact item/service cannot be bought at another retailer, and this means that we need to work harder to convince the customer that it’s our offering that’s the perfect fit. Asking relevant questions helps with that.?

Unlike a can of Coke that will be the same whether you buy it at a Walmart or your local gas station, a personalized grocery subscription service is vastly different from other ways of buying groceries. Prospective customer needs to be convinced that what we’ll give them will be a great fit.?

2. We are selling a subscription; each purchase is a fairly large dollar amount. Consumers, generally, resist subscriptions . Our average order value is $125, more than Instacart, more than an average grocery store trip, and a lot more than a single Doordash restaurant delivery. Both of these factors create a significant barrier to new customer signups. Tackling these, even at the expense of increasing the time it takes to sign up, produces a positive outcome.?

3. We are selling a service with Personalization as a key differentiator. Many consumer brands utterly failed to deliver on this promise, so consumers have become deaf to it. From Netflix suggestions that are just “popular movies in your zipcode” to Amazon’s product recommendations that are, again, just “popular recently,” consumers are skeptical . The privacy dilemmas - what happens to the data due to inherent incentive misalignment - exacerbate the mistrust .

How do we convince the customer that we’re different, given all this noise? By asking questions during the signup process that they would judge as relevant to delivering on the promise. How would a human chef onboard a new customer? By asking smart questions. In some way, when we “interview” the consumer in the signup funnel, they are interviewing us: by carefully observing what we ask.

Is your business a high-AOV, long-purchase-cycle business? Is your product highly differentiated, with no close alternatives that the customer can easily compare it to? If so, you may want to A/B test increasing friction in your signup funnel. You just might see a conversion improvement as a result.?