What’s a good conversion rate?
I get this question a lot, and not just about conversion rates. I always say “better than what you have now.” That earns me a glare that could sterilize cockroaches.
At the same time, I shove toothpicks into my toes to dull the sharp pain this question causes in my frontal lobe.
But that’s not fair. I know why people ask. If your conversion rate is 1/2 the industry average, and your bounce rate is 50% better, you want to focus on conversion rates. It’s hard to set priorities without benchmarks.
Here are three approaches. Use them wisely. Taken out of context, they can cause disasters that may get me punched in the nards.
First, there are some standards. If the corresponding metric falls below these, my gut says you need to work on it first:
Site load time: Over 3 seconds is bad because the average user just won’t wait any longer. Over 5 seconds is worse. Over 10 seconds means you drop everything you’re doing until you fix the problem.
Bounce rate: Sorry, there is no standard. Portent has an 85% bounce rate, and that’s fine for us. For you, a 10% bounce rate might be better. I’d use relative comparisons, instead (see below).
Time on page: Same as bounce rate.
Conversion rate: Under .7% is pretty ghastly. If you have a long conversion funnel, first-step conversions less than .7% are a problem.
I figured this out with real math: 90% of clients with a conversion rate less than .7% fire us or go out of business. That’s a 90% un-conversion rate. So it seems reasonable.
And now the hate mail starts. IANYOUIDIOTYOUJUSTGAVEANUMBERWTF. I know. I know. I’m sorry. Find me at Mozcon and tell me I’m a dork.
Skip this if you can. Most industry benchmark data is slightly better – only slightly – than guessing.
Google Analytics has audience benchmarking. You have to agree to share your data in the benchmarking pool. It’s all anonymous – competitors won’t see your site performance. Once you do that, go to Audience >> Benchmarking. You can compare channels, locations, and devices. And you can see bounce rate and session duration across each channel/device.
Site performance is easier. Go to a tool like Pingdom and test as many sites as you like. But every speed test tool measures different things. Use several different tools, and get a filmstrip view from a tool like Webpagetest.org.
Conversion rates are nearly impossible. You can search for and find all sorts of unverified studies based on agency data (cough). But they’re tiny data samples. Wordstream has decent data over here. They have a huge sample set that looks great, but their sample set includes (I’m sure) some serious marketing nerds. So their numbers may bias higher. Use with care.
Conversion rates, 2: This benchmark study on Moz.com — looks good at first glance, but it’s for only 30 sites. Again, use with care.
This is where you start. Because no matter how your competitors perform, improving your site matters more.
Before you worry about industry benchmarks, compare performance metrics across your site. Fix the negative outliers.
On Portent’s site, we’ll sometimes have a page with 90%+ bounce rate and sub–1-minute dwell time. The average is about 85% bounce rate and well over 3 minutes dwell time. The page in question is a negative outlier and needs a careful look.
Say your site has subscription forms with a 3% average conversion rate. If one form converts at 1%, that’s a problem. If another converts at 10%, something’s working well there and deserves investigation.
Make sure you’re comparing apples to apples: Visit/conversion volumes should be roughly the same. If one form got 10,000 visits, and the other got 100, you can’t compare.
Relative comparison is your best bet. But you have to get out of the what’s-a-good-conversion-rate mindset. That’s hard to do, I know. Just… try, OK?
The Bottom Line Is Your Bottom Line
What’s a “good” conversion rate? Better than you have now (Ian ducks).
Remember that you’re better off improving your site than chasing someone else’s.
I’ll to catch hell for this post. Truly. Send strongly-worded tweets to @realdonaldtrump. He’ll send ’em along, I’m sure.