Will This One Trick Boost Conversions by 400%? Probably Not: 6 Split Testing Myths Busted
- A split test with fewer than 100 conversions, undertaken for less than 2 weeks, is not truly statistically significant.
- Just because some guy on the Internet increased their lead generation by 143% with 2 more forms in their field, doesn’t mean you will, too.
- Conversion Rate is not the only significant Key Performance Indicator: so are traffic, Cost-per-Acquisition, Search Engine Ranking Position.
Split testing is great and can really make (or break) your online business presence. That being said, it’s not at all uncommon for an Internet marketing Singapore expert to claim they know all about A/B tests… after having read a couple of guides and blog posts on it.
Conversely, even seasoned pros, frustrated with not getting the results they expected from their latest conversion optimization campaign, can fall prey to some ‘quick fix’ type hack. False trends, myths and misconceptions abound in this world of highly sophisticated data analysis.
Today, we try to elucidate some of the most common myths about A/B testing, in an attempt to make your life as an optimization and Internet marketing Singapore pro far easier.
1. “Split tests end when they’re statistically significant.”
Ok, we’ll give you this much: this one is largely true. However, it’s not true for all A/B tests. According to Neil Patel, your best bet is to benchmark all of them for a minimum of 100 conversions for each variable. The timeframe the expert suggests sits at roughly 2 weeks.
Image source: Offer Vault
The truth: You can’t possibly determine if you have a winning variation or a fluke on your hands with less than 100 conversions, all amassed in a couple of days. In the long run, such flimsy results will turn out false.
2. “If it worked for them, it has to work for me!”
In the world of conversion optimization, ‘it’ can be any number of fine-tuning details. The specifics, be it a call-to-action button, social sharing icon, or security badge, don’t really matter. Don’t fall into the trap of assuming your online brand can emulate the test results obtained by another, in a completely different context (industry, niche, etc.).
Image source: Joao Paulo Pereira
The truth: Before you fall into the trap of the false positive, test, test, test. Draw your own conclusions.
3. “If it matters to them, it has to matter to me, too!”
This point is very similar to the one above. Perhaps you come across this test result, which says some site improved conversions by 143% just by removing 2 fields in their lead gen form. You resolve that the number of lead gen form fields is essential and implement the change. Meanwhile, your website wireframe crashes and is very buggy. Your conversion optimization efforts tank.
The truth: Don’t even bother focusing on the details, until you’ve made sure the bigger picture is as clear and smooth-running as possible.
4. “I just need to run one more test… And then another. And another…”
Remember when we told you above that test, test, testing is always a good idea? In fact, it’s not—not if you’re simply testing for the sake of gathering data, or trying to fix things that aren’t broken. More data, used poorly, is just as bad as no data at all, only more confusing.
Image source: Web SEO
The truth: Don’t test for testing’s sake. Focus on improvements which would really improve the user experience design and website accessibility.
5. “If only I had a good UX designer/engineer/information architect to work with…”
Oh, the things you would do then, right? Wrong. You see, what all of the above roles fill is the visual and end-user experience side of the deal. Designers and engineers are there to make sites look good and run better. Conversion optimizers and Internet marketing Singapore experts are the ones supposed to boost conversion rates.
The truth: While conversion rate optimization is a team effort, the person in charge of selecting design elements and website structure components to test for—is you.
6. “The conversion rate is the only relevant KPI.”
Don’t get me wrong, conversion rates are extremely important. But what about all the other performance indicators, which both online and offline business developers keep an eye on? What if your Customer Acquisition Costs end up running you into the ground? And what about Search Engine Page Ranks, Lifetime Value of an asset, revenue, and all that?
The truth: Key Performance Indicators, conversion rate included, are all delicately interconnected. Their relevance will vary according to your market/industry segment.