Nonprofit Website Benchmarks

By Brett

Your website is a unique snowflake with singular requirements. While there’s surely overlap, the make-up of your audience is different from every other site, because your mission – and the content you put on your website to support that mission – is different from every other nonprofit.

But you’ve probably wondered how your site compares to others. I know I’ve always wanted access to website benchmarks – some way to see if the trends I notice in our own dashboards are reflective of larger patterns.

Of course, benchmarks can be dangerous. If you obsess about them, the action becomes whatever the opposite of navel gazing is.

When you analyze the efficacy of your own website, you always need to consider it within the context of your own target audiences and organizational goals instead of worrying that your overall bounce rate is 4% higher than the “average.”

After all, if your site has the profile of a news organization, with links to recent articles shared widely through third-party channels, you should expect a higher bounce rate: people come, consume, and leave. That might be okay! The goal of that content may have been to spread some news, not generate donations.

Benchmarks can inform us about larger trends, though, and it’s damned annoying that they’re so hard to find.

It’s been almost four years since Groundwire published its 2010 Website Benchmarks Report. (Fortunately, you can still download the PDF from a third-party site.) Since then, Google has discontinued its Analytics Benchmarks Report – and even the semi-useful newsletter that followed.

You can order a $395 report from Marketing Sherpas (PDF), which may or may not be helpful; I certainly didn’t buy it. KISSmetrics published an infographic about bounce rates a couple of years ago with some interesting data… that may have come from a 2006 post to a Yahoo group.

Pew Research recently released an interesting study comparing news sites, using aggregated data from ComScore. I love how they’ve analyzed visitor loyalty across segments, but rare is the nonprofit that’s going to share the content goals of CNN or NBC News.

On our side, there are a couple of stats of interest in the latest Benchmarks Study from M+R and NTEN and the 2013 Online Marketing Benchmark Study for Nonprofits from Blackbaud, but they’re mostly focused on traffic growth and donation page conversions.

At ThinkShout, because we start almost all of our engagements by exploring how our client’s audiences use their current website, we felt it would be important to have some point of comparison. To facilitate discussion around questions like “How much do we need to worry about mobile?” or “How can we convert the influx of new visitors into engaged users?”, we’ve aggregated the data we have available to us.

This data is in no way reflective of the industry as whole, and it is very top-level. It represents fifteen organizations with diverse missions and traffic patterns, ranging from a few thousand sessions per month to more than 100,000. These are also generally organizations that have recognized the need to redesign their website.

With those caveats, we hope this data may help you understand that some of what you see in your own analytics may, in fact, be reflective of broader trends.

The following are three I see in the 37,000,000 pageviews we have access to.

Search Is the Ultimate Shortcut###

Back in 2010, Groundwire found that search engines referred 55% of traffic to the nonprofit websites in their study. While we can’t do a direct comparison since the sites in question aren’t the same, I feel comfortable making the blanket statement that the trend is toward more traffic coming from search.

Here are the mean / median numbers since 2011:

benchmark1.png
  • 2011: 47.06% / 43.97%

  • 2012: 48.12% / 50.12%

  • 2013: 52.99% / 54.02%

So far in 2014, those numbers have increased to 55% (mean) / 60% (median).

That means, of course, that traffic from the other two legs of the standard triumvirate have dropped:

benchmark2.png
  • 2011: 22.04% / 22.13%

  • 2012: 23.08% / 20.28%

  • 2013: 18.67% / 17.00%

  • 2014 (to date): 16.18% / 13.98%

benchmark3.png
  • 2011: 28.87% / 24.76%

  • 2012: 26.81% / 20.75%

  • 2013: 25.36% / 21.42%

  • 2014 (to date): 25.82% / 19.75%

Essentially, the best home page ever created will not solve your problems, as your users are more and more likely to turn to search first to find what they’re looking for. Your information architecture must take into account the fact that the first page your visitors encounter may be deep in your website.

User Experience starts in the first place your users experience you. You can use your data to make some assumptions about where that’s mostly likely to happen and optimize the top landing pages, but, more and more, you need to worry about every piece of content on your site.

Stop Wondering if Mobile Is Important Right Now###

It is. Groundwire found that in 2010, the median number of mobile visitors to the sites in their study was just 1%. That’s changed over the past few years. Looking just at mobile phones (not tablets), you can see the surge:

benchmark4.png
  • 2011: 5.70% / 4.97%

  • 2012: 7.68% / 7.67%

  • 2013: 12.84% / 12.04%

  • 2014 (to date): 18.68% / 17.45%

I’ve seen the argument made that mobile doesn’t require your attention quite yet because it still represents, according to this data, just 1 in 5 visits.

Our suggestion would be to get ahead of the curve instead of fighting to catch up. The growth in mobile traffic is having an obvious impact on the most basic measures of engagement.

 Mean  Median  
YearBounce RateTime per SessionPages per SessionBounce RateTime per SessionPages per Session
201162.30%2:042.3067.07%1:572.09
201266.17%1:582.0667.20%1:561.89
201370.48%1:541.8770.10%1:391.84

[I’ve excluded 2014 because I haven’t cleaned out the data from sites that have gone through a responsive redesign. This isn’t meant to be a marketing piece, but I can tell you that redesigning for mobile can have a significant impact.]

My suspicion is that as people grow more comfortable using their smartphones to browse the Internet, visiting brands that have implemented a mobile strategy, they’re less likely to put up with sites that don’t work as well as they expect in their mobile browsers.

By structuring your content properly, you can create ways to put the content mobile users are most interested in front and center. If nothing else, you should take the time to understand what mobile visitors are doing on your site as a first step.

Loyalty Is Hard to Come By (or Even Define)###

Perhaps, given the growth in search and mobile traffic, it’s no surprise that the percentage of “New” visitors has increased over the years:

benchmark5.png
  • 2011: 65.65% / 66.01%

  • 2012: 70.34% / 70.13%

  • 2013: 74.35% / 75.00%

On the flip side, “Loyal” visitors (defined as those with at least three visits in the period under review) have crashed:

benchmark6.png
  • 2011: 20.86% / 19.00%

  • 2012: 15.64% / 15.66%

  • 2013: 12.16% / 12.26%

And those numbers aren’t just relative percentage drops, caused by increases in other types of traffic, because the real numbers have dropped as well:

  • 2011: 44,734 / 33,894

  • 2012: 43,076 / 29,989

  • 2013: 37,757 / 22,781

I’m fairly confident, given the scale of the data, that this isn’t simply because people are getting better at clearing their cookies regularly – there’s wide variation in estimates, much of it from surveys of user behavior, rather than actual data – or browsing the web incognito. Those technologies have been around for years, and I doubt even Mr. Snowden has had that much of an impact on people’s everyday habits. (But I’m willing to be convinced otherwise.)

This loss could also be reflective of device fragmentation. Google and others are working on ways to track users across the many devices they use, but right now, if somebody visits you on August 14th on their laptop, comes back on September 9th on their tablet, and then again on October 21st on their phone, they would be classified as three different New users. As we see mobile traffic grow, it may mean that right now, we don’t have an accurate way to track visitor loyalty.

Perhaps some of that loyalty is being transferred elsewhere, however, from the website to social spaces, third party donation pages, or even mobile apps. In that case, the better sites are at converting website visitors into these different kinds of traffic, the less relevant the website will become to them.

Rare is the constituent who comes once a year to donate. Engagement, even within a multi-channel strategy, has to strike a balance between assuming that visitors will find continued value from the resources we make available on the web and moving new visitors to spaces where they’re more likely to continue to interact with us and each other.

In any case, if visitors are increasingly less likely to return to nonprofit websites, we need to rethink some of our engagement strategies. If they’re turning elsewhere because they aren’t finding what they want, when they want it, we’ve got some serious work to do.

Interested in Helping Out?###

I’ve made the spreadsheet with the cleaned data for 2011-2013 publicly available. We hope you’ll check it out and add insights of your own!

But beyond using it as a tool to inform your own website, we could use your help. What other data should we track and share? What do you want to know?

And more importantly, are you willing to share your own?

If you’d like to work with us on a template that can be used to collect data from other nonprofits, just let me know. We’ll keep the org-by-org breakdown anonymous. Once we reach an arbitrarily large threshold – 100,000,000 pageviews? – we can do some follow-up work.

Lev Tsypin, ThinkShout’s Director of Engineering, has suggested that we could go so far as to set up a Google form that feeds an anonymized spreadsheet. We could then make that data available via JSON, for further manipulation.

In the end, I believe that the more we aggregate and share, the better we’ll identify the problem areas we need to focus on as a community.