The Case Against A/B Testing at Early-Stage Startups
If you’ve been listening to the tech startup illuminati over the past few years, you’d be forgiven for thinking that the cult of A/B testing has become permanently ingrained in the “lean startup” mentality. Indeed, whenever there is a product question that cannot be answered by mere intuition, today’s rock star startups seem to unquestionably advocate the so-called “split test” as the arbiter of all crowd wisdom.
A/B tests have become so ubiquitous that founders who don’t frequently employ them have been subtly made to feel bad about themselves for having such a weak data-driven culture. This has gotten a little out of hand. Yes, it is true that A/B tests can provide tremendous insight into online user behavior, which can -- with a lot of data -- help you improve your conversion rates. But there is a problem.
The problem at early-stage companies is that early optimization can distract management from much higher yield activities, while often not having a large enough sample size to yield meaningful results.
Performing split tests is simply a management-intensive task, no matter how cheap and efficient A/B testing plugins such as Optimizely have become. Someone needs to devote their time to determining what to test, setting up the test and verifying and implementing the results of the test. Even if this technically does not require very much “time,” it requires management’s mental bandwidth, which is the scarcest resource at any company (especially an early-stage startup).
Do you really want your co-founder to be distracted by yielding a 5 percent better conversion rate (if he or she is lucky), when you haven’t yet proved that there is a burning need for your product or that you are able to scale your marketing effectively?
Similarly, how can you be sure that those A/B test “results” are statistically significant in the first place? Young companies typically have low amounts of web traffic, which may not produce usable A/B test conclusions for several weeks or months, if at all. And the frequently changing nature of your web traffic can make it very hard to draw meaningful conclusions to begin with. If 80 percent of your traffic this month came from a single TechCrunch article or Facebook ad blitz, is that really representative of how your core user base will behave on that target web page in the future?
Early-stage startups should therefore be careful not to get too bogged down in perfecting their marketing copy or the color of their call-to-action buttons until they are confidently nailing all the fundamentals. Building a metrics-driven culture at your startup means ensuring that you are always focusing on the right key performance indicators and not optimizing them too early. Sometimes a quick focus group, hallway usability test or intuition can be sufficient for you to make product decisions that are “good enough” for your company’s current size.
Have you had positive or negative experiences with A/B testing at your startup? Sound off in the comments.
Entrepreneur Editors' Picks
Kale Was a Garnish Before This Creative Genius Made It Famous. Here's How She Did It — and What She's Planning Next.
Telling Your Brand Story Is Crucial. 4 Steps to Ensure That It Resonates.
This Baker Was Told Not to Speak Spanish With Colleagues, So She Started Her Own Cake Company That Values Employees Just as Much as Customers
Improving Yourself Takes 9.6 Minutes of Work Each Day
Meet the Women Behind Some of McDonald's Most Iconic (and Essential) Ingredients — and How They're Setting New Standards
Remote Work Shouldn't Be Up for Debate
Employees Are Over Foosball Tables and Free Snacks. Your Company Culture Needs This Instead.