Let’s forget about this year’s presidential campaign because I’ve run out of popcorn.
And let’s pretend it’s the year 2007 and you’re managing Obama’s campaign.
He asks you to help him gain greater support online and you are thinking about using one of the following:
Take a moment and give your best shot. I will reveal the answer to you in a bit.
Now, here’s the deal: If you’re guessing right now since it’s a 50/50 chance for each option, I suspect that you won’t stay in this job for long. It’s the nation’s future at the stake and you’re just winging it!
On the other hand, if you’re thinking “I don’t know which one will work better, but I can test it”, you nailed it.
It turns out, a video performed 30.3% worse than an image of Obama. And that little test contributed 4 million of the 13 million addresses in the campaign’s email list and resulted in $75 million in funding. (Source: The A/B Test: Inside the Technology That’s Changing the Rules of Business)
My point: Your instincts can only get you so far. Don’t guess it, test it.
In a world of testing and data
You don’t always have a presidential campaign to manage, but chances are, you might be considering adding a new feature to your product, or wondering if a new website performs better than your current one. For those reasons, doing A/B testing will help you make better decisions.
Example #1: Which name should I use for my new product?
For the last 2 months, three business owners asked me “What do you think of this name for my new product?”
While I feel flattered that they trust me enough to ask for my opinions, I have never stated one. Instead, I always ask them “Why don’t you run a test? Let data tell you.”
It’s a simple test you can set up right now.
If you have a LinkedIn/Facebook group that’s filled with your prospects, list all the potential product names and do a poll. Let the number of vote tell you which name resonates with them.
Or, you can set up a Google Adwords campaign and let the number of clicks tell you what name is the most popular one.
You don’t need to or shouldn’t make a decision based on someone’s opinion. After all, if you can test everything, why would you make decisions based on guesses or luck?
Example #2 Which offer can convert more website visitors to customers: 30 day free trial or full-price offer?
Let’s say that you’re the owner of a company called Metflix.
You want to convert more website visitors into paying customers, and you are considering one of the two options:
You wonder which one can convert more visitors to full-price customers?
So you run a A/B test, it might look like this:
You might think 40% is obviously better than 20%, so you should go with the trial offer.
What if I tell you that after 30-day trial, only 55% of prospects choose to continue on a full-price basis?
Now it looks like this:
It gives you some idea how this works.
Of course, you may also want to take it one step further. For example, after being a paying customer, how long does each group of customers stick around?
There are two constants in A/B Testing: a scientific process and math
A/B Testing doesn’t need to be complicated (although it can be). Testing velocity varies, but the fundamentals are the same:
Summary: Test it like your life depends on it.
Sign up for my newsletter below and never miss a post like this.