The Three Metrics You Need to Know Before You Waste Any Time on A/B Tests
It’s hard to argue that split testing (also know as A/B testing) is changing the face of marketing. According to Unbounce, 44% of online businesses are using split test software. And software products like Unbounce and Visual Website Optimizer are making it ever easier. Split testing, done right, with good context, can put a stop to all the guesswork, anecdotal conclusions, and correlation/causation errors that can abound in marketing circles.
But it’s not without risks: split tests are expensive to run, requiring investment for both software, and staff/consultants to run the tests. Not to mention the opportunity cost of lost time exploiting other profit levers in your business.
All of which underscores the importance of testing the right metrics in your business, and the potential cost in time and resources of testing the wrong ones.
While I can’t speak for all businesses, what I’ve seen again-and-again with clients and peers is businesses gravitating toward what’s easy to test – landing pages, checkout pages, email subject lines, and sales pages (all of which can be extremely important in the right context) – rather than what’s important.
That’s why one of the most meaningful changes you can make in your business is to implement a process for identifying which parameters to test and optimize. Below are 3 metrics you need to know before you spend one more minute split testing.
1. List-to-Sale Conversion Rate
What if I told you one simple calculation would tell you whether to optimize any conversion metrics between an opt-in and a sale, or to look elsewhere? That’s what the list-to-sale benchmark gives you. “List-to-sale” is the percentage of buyers of your product or service over a given time period relative to the number of opt-ins to your email list for the same period.
Say in a given month you get 1,000 opt-ins to your email list, and in that same month, you make 55 sales of your flagship product. Wondering whether you should go with a webinar funnel instead of an email onboarding sequence? Whether to incorporate video into your sales page? Whether to change the color of your “buy now” button?
The answer to all of them is “no”, and I didn’t even need to take a look inside your funnel. Why? With 55 sales, you’re converting at a staggering 5% list-to-sale.
To calculate, just take the sales in the last 30 days and divide those by opt-ins over the same time period.
Some readers will be noticing the absence of a sales cycle in that calculation (i.e. since it takes days-to-weeks and several touch points to make a sale. We should be comparing this month’s buyers to last month’s opt-ins). You can control for this with a simple average:
- Take the last 4 months, and average the opt-ins over the first 3
- Then average the sales over the last 3.
- Then perform the same percentage calculation. (Sales divided by opt-ins)
For example, say you’re calculating in August:
- First you’d average the monthly opt-ins for April, May, and June. Let’s just say the average is 1500.
- Then you’d average the sales from May, June, and July, in order to leave a 30-day lag. Let’s say that average came out to 75.
- Dividing the sales by the opt-ins, and you’d get 5%.
The benchmark you should be aiming for? 1-2%. Below that, go nuts with split testing parts of your funnel. Above 1%, look elsewhere.
Above 2%, and I’d seriously consider raising your prices. In the hypothetical case of the 5% from above, I’d immediately double the price.
Next, and especially if your list-to-sale conversion is at-or-above the 1-2% benchmark, it’s time to look at your traffic.
2. Opt-in Conversion Rate
The vast majority of businesses I work with have list-to-sale conversion rates closer to benchmarks than their opt-in conversion rates. Put another way, if they’re wasting any resources split-testing their funnel or sales copy, they’re completely ignoring the sizable cohort of website visitors who never even see the offer because they bounce off the site.
As with list-to-sale conversions, you can do a back-of-the-napkin calculation for opt-ins. Just count your new subscribes from the last 30 days and divide it by total website visitors during that same 30 days.
The benchmark to aim at for opt-in conversion is 10%.
If you haven’t ever found your opt-in rate before, my guess is you’ll be astonished how low it is. I’ve seen it as low as 1-2%.
Luckily, there’s a simple strategy to improve it:
- Find the individual opt-in rates of your biggest webpages and your 10 most popular content pieces. (If you’re using a plugin like SumoMe or OptinMonster, you can set up the software to tell you your opt-ins for each page.)
- Look for the “outliers” – content pages often perform worse than home and about pages.
Once you’ve identified the worst-performers, perform this simple checklist (from lowest-hanging-fruit to more subtle)
- Can readers find your opt-in offer, or is it buried below the fold or ¾ of the way down a blog post?
- Are you giving your visitors only one thing to do on each page or post, or are you offering 3 different giveaways on various parts of your page?
- Is your opt-in offer not just well written, but well copywritten? Does it specify exactly who it’s for, describe a clear, specific benefit, and emphasize the urgency for opting in? (Even high performing opt-ins can usually be improved).
- Are you requiring your subscribers to double-opt-in? This will lower your opt-in conversions. Many founders I’ve talked to like to use a double-opt-in because it seems more “polite”. In my opinion, making somebody go off the page to get the freebie they just gave your email address for, let-alone to wait up to 20 minutes for it to arrive in their mailbox isn’t particularly polite. When I give my email address to get a lead magnet, I want it now – not after reconfirming my email address and waiting 20 minutes for the email.
This is not the type of page you want to create if you’re looking to increase opt-ins.
Split-test ninjas take-note: if you’ve read this far, and your opt-in rate is indeed garbage, there’s ample opportunity to split test:
- Two versions of a homepage with different opt-in copy/design.
- Two versions of an exit-pop on a popular content piece.
If you’re among the extremely lucky minority with list-to-sale conversions at-or-above 2%, and opt-in conversions at-or-above 10%, and you’ve raised your prices, I have some disappointing (although kind of good) news: split testing is not a good fit for your business.
Here’s the question to ask: Are your monthly sessions at least 50% of your list size? (i.e. if your list has 2,000 subscribers, are you getting at least 1,000 uniques per month?) If not, you need a traffic strategy. Don’t waste your time A/B testing anything.
While I’m a conversions expert and not a traffic expert, here’s a quick decision tree:
- Determine your market size. If you could 5x your traffic, are there enough people in your market to support it?
- Implement a content/syndication/guest-post strategy ASAP. It’s practically the only guaranteed winner across all verticals, but it can take up to a year to bear fruit.
- Consider hiring a paid traffic expert for one month to test customer acquisition costs from various paid sources. Choose the most profitable and double down while you wait for organic traffic to grow.
Bottom line: the same month spent split testing two opt-in offers on a homepage, landing page, or content page, could provide a 2-4x increase in revenue (by, say, improving an opt-in conversion rate from 1% to 4%), while the same time and money spent trying to boost an already maxed-out sales conversion rate would have a much smaller return.
That’s why a little context can save you thousands.
About the Author: Nate Smith is a direct-response copywriter and funnel expert who helps businesses scale by exploiting their most powerful profit levers. Nate is founder of 8020MarketingGuy.com.