Dayspring Technologies

It’s been just about 5 months since we re-launched the Dayspring website with a fresh design and a custom content management system. We’ve gotten quite a few positive comments on the site’s clean, precise look.

But how has the site performed? Sure it’s a laudable goal to make the Web more beautiful, but, frankly, most sites (yours and ours) are meant to do something. The “something” in our case, is to make sure that folks who want more information about Dayspring can get it.

“How your website has performed?” you ask. “Can you really say anything concrete about that?” Actually, yes. Two key numbers: +61% and +38%.

First, the Source Data

Before we can get to the answers to the question, we need to have some data to work with. That’s why analytics should be a key part of any redesign. We’ve been running Google Analytics (GA) on the Dayspring site for a couple of years now (we were on DeepMetrix’s LiveStats before that), and so we have a good baseline of data to compare against. We set it up to specifically exclude traffic from Dayspring’s office by creating particular Profiles within GA that apply various Filters.

Top-Level Statistics

Here are some callout statistics, showing our improvement on two key metrics:

• Average Time on Site. This is the average length (in minutes and seconds) of each visit to our site. Multiple visits by a single visitor are each counted. Before the redesign we were averaging 1:47 per visit. After the redesign, this figure increased to 2:12, or a 60.7% improvement.
• Average Pages per Visit. This is the average number of pages each visitor views during the course of a visit to the site. Again, multiple visits by a single visitor are each counted. Before the redesign, we averaged 2.77 pages per visit. This figure increased to 3.81 after the redesign, a 37.5% increase. (See Fig. 1)

Note that both of these are valid figures independent of how many visitors are coming to our site, and they are a good measure of the engagement of visitors. On average, visitors appear to be much more engaged with our site: viewing more pages with each visit and spending more time.

Maybe It’s Just Because…

Of course any good analyst will ask whether there are other factors that may be in play.

Is the average time on site higher because the site is slower—resulting in a negative user experience on balance?

No, because Google’s Webmaster Tools is actually reporting that our site is very responsive. Although you can’t see it here, the average of 132 ms is lower than prior to the re-launch.

Is the user clicking around more because there is less content on each page?

No, in general that wasn’t a key feature of our redesign, although we have done a better job accommodating readers who scan with a multi-level visual hierarchy.

Is it because we went out and told all our friends that we launched a new site and they came and spent hours and hours on the site?

That’s an interesting one. Remember that we already filtered out visits that were from our own office. Not doing that is a great way to pump up your stats, but it’s also incredibly misleading. To work to eliminate “friends and family”, we set up some Advanced Segments in GA to only look at visitors who a) came via search and b) didn’t use “dayspring” in their search keywords. Not perfect, but it gets you pretty close. The stats, as expected, were not identical, but they still showed very significant increases in both average pageviews and average time on site per visitor.

A few other statistics also confirm that the site is performing better than our previous one (which, rest its soul, served us well for 3 years):

• More people returning. There were 10% fewer visitors in the “only visited once” category after the redesign. And more in every other category. There was a 21% increase in people visiting twice, 53% in those visiting 3 times, etc.
• More people spending more time, fewer spending less. This seems like a corollary of average time on site being higher. However, this point goes to the distribution of those visitors. GA segregates visits into 0-10 seconds, 11-30, 31-60, 61-180, 181-600, 601-1800 and 1800+ (more than 30 minutes). For visits after the redesign compared with those before, all visits in categories less than 60 seconds are down, and visits in categories more than 60 seconds are uniformly up. This is exactly the behavior we would want to see and tells us that our average data is not just a statistical blip.

What About Conversions?

Conversions are usually the gold standard for a website. Every site has a particular reason or reasons for existence. That reason may be to disperse information, build community, connect mentors and entrepreneurs, make sales, garner donations or connect with people interested in services. When the desired action happens, we call it a conversion and we can measure these. In our case, we measure when people submit their information via the form on our website.

If people are visiting our website, are they actually contacting us?

Interestingly enough, our conversions are about even with the comparison period. A slightly greater percentage of our visitors are visiting our Contact Us page. But a slightly smaller percentage are actually submitting the form. We haven’t changed very much about that page, so we’ll have to look into why that might be the case. A few possible reasons:

• As part of the redesign we renamed the page “Get an Estimate”. As a result, folks may be expecting a project cost calculator (which we don’t have—we prefer to talk to people because of the type of projects we take on).
• More people are calling us instead of submitting the Web form because we have made our phone number more prominent in the “Ready to start?” widget on each page. Total overall leads are, in fact, up.
• Economic conditions have more people doing research but fewer ready to actually have a discussion about a project.

Conclusions: What Worked? And What Didn’t?

Matt Crawford, our resident statistician (and PHP code jockey), tells me that it’s not possible to say whether these improvements are statistically significant in the strict sense. (You would need standard deviation information in order to make that pronouncement—information that Google doesn’t supply. Just imagine if one user decided to click 3 million pages in one session. That would skew your average pageviews enormously). However, I think it’s safe to say that the design, just on the numbers, was successful.

So what worked?

When we approached the redesign we had a few key goals with respect to usability:

• Be kinder to scanning readers
• Make it easy for people to contact us
• Promote the ability to navigate fluidly
• Make the site fast

And so we made sure that our site incorporated a few key features.

• Visual hierarchy. We make sure that there is as solid visual hierarchy (e.g., three well-differentiated text sizes) to allow people to scan if that’s what they have time for. (Fig. 3)
  
• Encourage conversions. We prominently include the “Ready to start?” widget on each page to encourage people to contact us. We also include our phone number in this widget for those that are more comfortable with picking up the phone and talking to a real, live person rather than sending a query into the Internet ether. Prior to the redesign, the “contact us” link simply wasn’t very prominent. (Fig. 3)
• Directed path. In general, we make sure that each page has an easy “directed path”, leading to another page of content—usually a link right after the main page content which uses our highlight color (orange). (Fig. 3)
• Surface related content. For our portfolio, we visually display other related work in the right hand column to encourage browsing. (Fig. 4)
• Technical optimizations. We “minified” our JavaScript files, paid attention to code ordering for optimal perceived rendering time, and followed a number of other best practices.

Along the way, we also solved some problems with the old site, including navigation that didn’t look like navigation. This, we believe, has led to the significant improvement in key measures of engagement.

So, what didn’t work?

While we’re quite pleased with the site overall, it still does have some problems. One major one that we uncovered in this analysis is that our statistics show that the “Capabilities”, “Work” and “Team” pages are not visited as much as they ought to be. In fact, the sub-pages “Team: People” and “Work: Our Clients” are each visited more than their respective “parents”.

We designed the parent pages to give an overview and have put important information on there to orient people to the respective sections. Why do they appear not to be getting the visibility we want them to have? We suspect it’s because of the way the dropdowns in our top navigation behave:

The dropdowns list the sub-pages but not the “parent” page and it appears that users are hovering over the parent and selecting one of the children—ignoring the parent page. For now, we’ve left this for you to see and play with for yourself, but we’ll be fixing it shortly.

Takeaways

Finally, some key takeaways:

• Put analytics on your site. For a properly structured site, analytics really don’t take that long to implement. So make sure you do it now. You can’t use data that you haven’t collected. Oh, and while you’re at it, set your stats up to include the ability to exclude your own visits. You know that your staff spends a lot of time on your site. What do other people do? That’s the more interesting information.
• Keep an activity log. Know the key dates in your website’s history. For example, record the date your redesign was launched, the date of any significant media mentions, the date you ran an experiment. This also helps you track any effects that changes you make have on search engine results.
• Keep your old site around for reference. If you’re able to, keep your old site around for reference for a little while. Or, even if you’re trying your hardest to forget that it ever existed, minimally keep some screen captures to remind you of it. You might want to do some comparative reporting or to understand how subtle changes you made affected user behavior.
• Setup conversion goals. Be clear about what you want your website to do, and instrument your site to track that. Your conversions could be a successful contact form submission, a successful sale, a download of a whitepaper, or any other visitor interaction that is important to you.
• Give them somewhere to go next. Try to do a little self-observation next time you’re browsing the web. When you get down to the bottom of a page where do you go next? If there isn’t a clear pathway for what’s next, what do you end up doing? Does it affect your behavior if there’s a big link there just waiting for you to click on it? If you’re like most people, the answer is probably “yes”. Give people somewhere to go when they hit the end. Keep them in a virtuous loop and on your site.
• Make your site fast. Performance has an impact on whether folks will browse. It’s worth it to spend the extra time to make your site as speedy as possible.

Do you want to see some more hard numbers? Leave a comment here and we’ll be in touch. Or, do you need help with analysis, strategy or a redesign? Get in touch with us.