Friday 29 May 2020

“What’s a good conversion rate for my landing page?” [New AI-Backed Research]

Imagine you’re an ecommerce business using landing pages to sell sweaters for miniature pigs. Recently, you’ve done a round of A/B testing—adding a little more oink to your calls to action, let’s say—and tweaked your social ad targeting to reach the hardcore piggy people on Instagram

After all your optimization efforts, your landing pages now convert at 3.57%

But actually…even if it’s a big improvement against your personal baseline, how do you know you should stop there? How do you know that your hard-earned conversion rate is worth celebrating? Heck, how would you even know if a 30% conversion rate is any good for pages in your industry? (Maybe everyone’s getting a fat return off of pig sweaters but you.)

It’s hard to be confident in the numbers when you don’t know how everybody else is doing. Doubt settles in. Maybe you’re missing out on reaching your conversion potential without even knowing it.

Well, we feel your pain. That’s why, at Unbounce, we’re on a continuing mission to answer the big question for you. It’s the one we hear time and again from our customers:

“What’s a good, bad, or average conversion rate for my landing pages?” 

That’s where industry benchmarks come in—and that’s why we’re thrilled to bring you a fresh (and free) new version of our Conversion Benchmark Report

Benchmarks can energize your digital marketing strategy in three big ways:

  1. They’re a form of competitive intelligence. They help you identify gaps between your performance and what the rest of your industry considers to be a good conversion rate. 
  2. Our benchmarks reveal data-supported best practices, and you’ll waste less time and traffic testing unproven optimizations that our machine learning analysis shows don’t necessarily work. 
  3. They help you build a culture of continuous improvement in your organization. It’s harder for your marketing team to be happy with “just okay” if they’re seeing something to strive for.

Sure, some folks like to pooh-pooh industry benchmarking—“Why should I care how other marketers are converting? Why don’t I just focus on how I’m doing?”—but they’re your best window into what success really looks like. Going forward blindly, when you could have both eyes on the prize, is just silly.

Oh, and these benchmarks were generated with help from an honest-to-goodness AI crunching millions of conversions, so the results are far more reliable than the anecdotal best practices often found online. As part of the Unbounce Conversion Intelligence™ approach to digital marketing, these machine-derived insights help you pair your hard-earned expertise with AI to create the highest-converting campaigns of your career.

Introducing the 2020 Conversion Benchmark Report

This year’s Conversion Benchmark Report uses machine learning to assist our data team in analyzing 186.9 million visits to 34,132 Unbounce-built landing pages. In terms of sample size, we analyzed more visits to these pages than the populations of Canada, Hong Kong, Mexico, Laos, and Ireland combined. 

For full context, the previous (2017) version of this report was also built on machine learning insights, but in three years we’ve refined our approach to provide tons more real, proprietary customer data to feed the machine. Now we have even better, more reliable outputs—as well as a few new ways to break down our findings, like by conversion goal. (These are boundaries we’re going to keep pushing, too.)

But what kind of info does the report contain? For one thing, you’ll find median conversion rates broken down to 16 key industries. In many cases, we’ve got wide enough sample sizes to sort them into subcategories too, so you can see how your brother’s pest control service measures up against your sister’s HVAC company. (Or how your uncle’s cybersecurity software converts against your great aunt’s cloud accounting platform.)

The Conversion Benchmark Report includes 16 industries broken down into dozens of subcategories.

Why do we report on median instead of average (mean)? Our goal is to provide you with a realistic picture of where you stand, so this year’s report lists median conversion rates as our measure of central tendency instead of the mean. We found this reduces the impact of outliers (like pages that convert five times better than the rest) on the final benchmarks.

Not clear enough? Then imagine you want to find out, on average, how many eyes people have. The median tells us they have two eyes. According to the mean, though, they have slightly less than two. Because outliers (people with one or fewer eyes) bring that number down.

Both these measures are correct, but which one would you prefer to rely on if your business is selling sunglasses?

What if your industry doesn’t appear in the report? For this year’s report, we’ve tried to be even more representative. With machine learning helping us to sort thousands of landing pages in a logical way, we’ve increased the number of industries covered from 10 to 16, and we’ve even added subcategories whenever sample sizes allow.

If you still don’t see yourself represented, though, compare your conversion rates to industries with similar audiences and conversion goals. While we don’t actually recommend comparisons between very unrelated industries (except for fun), let your judgment be your guide. 

A note on COVID-19. The conversion data in this report comes from 2019, so we realize it shows norms that have been disrupted for some vulnerable industries—like travel, events and leisure, restaurants, and medical practitioners. These benchmarks show what you can expect in stable periods, and they provide insights about how your visitors typically behave (and why they convert). We hope they’ll help you set up your digital campaigns for success—and inspire your rebound.

If you face uncertainty, though, please also check out the COVID-19 Small Business Care Package for a roundup of useful resources to help lessen the impact on your business.

Below, I go into more detail about the findings and insights we’ve been able to pull from them. But if you’ve got an itchy mouse-finger, you can jump right into the Conversion Benchmark Report now. (It’ll open in a new tab.)

Going beyond the benchmarks

Benchmarks are tremendously helpful, for all the reasons I talked about above. (If you work for an agency, you know this already. They’re a baller way of showing the value of what you do—and helping clients determine their true conversion potential.) 

How do I best communicate with my target audience? 

In copywriting circles, the received wisdom is that clarity comes above all else. If you’re looking to put up the fewest hurdles possible between audience and offer, it can make sense to keep your vocabulary basic and your sentences tight and untangled.

Our data, however, complicates this equation. Is simple always better? Nope. It turns out that different industries tend to convert more often at different reading levels (and some see weaker relationships between conversion rates and readability than others). 

There are even cases in which it’s good to sound sophisticated. B2B companies offering lead-gen consulting or instruction, for instance, appear to benefit from more challenging language. We see a drop in conversion rates as pages become easier to understand. (Frankly, that’s not what we expected.)

When it comes to reading ease, pages for lead-gen consultants appear to benefit from being harder to read.

Our machine learning analysis enabled us to look at copy from 34 thousand pages. Each page is assigned a Flesch reading ease score based on the average number of syllables per word and words per sentence. More syllables and more words means more…harder.

Here’s roughly how the scoring breaks down:

What’s the perf word length for my landing pages?

While it’s true that shorter pages tend to convert better, many industries have sweet spots that break the rule—which means, if you’re going to create a long-form landing page, you should go this long. This is especially true in the wild territories beyond 200 words, where unexpected correlations between length and conversion rate have led many a marketer astray.

At what length do landing pages for family services convert best? The graph provides answers.

Depending on your offer and industry, you may find that you need to use more words to get your point across, but graphs like the one above can let you know what’s ideal. For family services, that’s 300-500 words (if you can’t get it shorter than 150 words). For other industries, it can be more or less. Whatever the case, creating variants based on our findings can definitely be a good candidate for A/B testing or Smart Traffic.

What emotions might relate to better conversion rates?

You likely know in your gut that people’s feelings can impact their decision to buy, but which ones actually drive conversions on your landing pages? To find out, we ran an ML-powered sentiment analysis that looked at emotion-associated words that might relate to healthy conversion rates—and which might even be slowing you down. 

(Spoiler: using trust words isn’t always advisable. “Trust us.”)

For SaaS, the concentration of anticipation words on a landing page correlates with its conversion rate.

When it comes to SaaS conversions, for instance, it turns out that language that conveys anticipation (words like gradual, highest, improve, and launch) sometimes correlates with better conversion rates. Or, to put it another way: as we find more of these words, we also often tend to see better conversion performance.

You can explore this example, and many others, in the report.

A good conversion rate is one you can improve upon.

When it comes down to brass tacks, all this benchmarking is valuable insofar as you can use it to build a better conversion machine from what you learn. How do you do it?

  1. Explore the insights in this report. The report is broken down into 16 industries. How are your landing pages stacking up against the baseline? Are you way out ahead? Are you falling behind? Start with your industry, sure, but take a look at others too. There may be insights that are worth exploring outside your own arena.
  2. Apply the data learnings to your own campaigns. Create a variant (or more than one variant) of your page that applies some of the insights we’ve provided. For example, you might dial down the jargon until you hit the optimal Flesch reading score. (You can use the free readability formula tool here to test it for yourself.)
  3. Optimize and test. Keep in mind that data analysis reveals trends and tendencies rather than absolutes. You’re making informed decisions when you apply these learnings, but testing is still your best way to confirm. Run A/B tests or, if you’re short on the time or traffic to do so, just publish your variants and turn on Smart Traffic in the Unbounce Builder. It’ll use machine learning to automatically decide which variant is right for which visitors, and it’s otherwise hands-off. (If you’re looking for more ideas on how to build variants, I’d recommend this post from Garrett too.)

In short, this year’s report uses ML to identify opportunities you simply couldn’t spot without the processing power of a machine. Optimizing your pages doesn’t have to be aspirational. We believe this is the future of digital marketing—and, going forward, you’re going to see more and more efforts like this from Unbounce to help you enhance the skills you already have. (If you’re curious about what else we have planned, you can read more about our push to bring you Conversion Intelligence.)

Whether you sell pig sweaters, chicken harnesses, or something altogether more practical—are you confident enough to swagger into your next meeting, snap your suspenders, fire those finger-guns in your boss’s direction, and let everyone know about your team’s big win? “Soooooooo-ie!”

Take a gander at the 2020 Conversion Benchmark Report, and let us know what you think in the comments below.



from Unbounce https://ift.tt/3gBecfe

No comments:

Post a Comment