optimizable: learn to optimize a webpage in 5 minutes

target

"If you can not measure it, you can not improve it."

Lord Kelvin

Every web page has a purpose. What's the user supposed to do when he lands on your page? Out of the gazillion pages on the web, somehow, he landed on your page. What a preciously rare opportunity. How will you capitalize on it? How does your content and design encourage the user to your preferred next action? Do you even know what your preferred next action or set of actions is? Do you measure them? Do you try new things to see how they affect the user's actions relative to your preferred actions?

You probably don't. But don't fret, almost nobody does. You can go out and read a bunch of books about conversion rate optimization, web analytics, and multivariate testing, or you can just read the rest of this page in five minutes and leave with a concrete plan to start optimizing your webpages today (you can read all the other books later, they're quite good, just not the best place to start).

Does your website work?

Is your website doing what it's supposed to be doing? Probably not. And definitely not as well as it could. Website testing and optimization is not easy, but let's make it easy enough so that you can start today. We'll look at two little known, but fundamental building blocks for page optimizaton – idle rate and traffic flow.

People doing nothing on your page is not success

What is your page supposed to do? How well is it doing it? Can you measure it? If you can't measure it, you can't optimize it. Here's where you can get lost in endless discussions around "what is success". Don't.

You usually want your page visitors to do something on the page. Buy something, view another page – click a button, click a link. So at the highest level, you can define success as the opposite of failure – you don't want people doing nothing. I call this your page's target idle rate.

Idle rate is defined as the percentage of page views in which the user does not click on anything on the main body of your page. If the sleepuser clicks on the global header, footer, or search, that does not count. Credit is only given if the user engages in the page content unique to the page. The lower the idle rate, the better. Idle rate is bounce rate's cuter, younger sibling.

Idle rate is your page's top level success criteria. It's one big number that you can measure against to see how content and design changes can favorably improve it.

If your page had 1,000 views and users during 250 views did not click on anything in the main body of the page, its idle rate would be 25%. If you changed the visual design of the links on this page and now only 150 views did not register a click on anything in the main body of the page, you've improved your page's idle rate to 15%. With a measure as simple as that, you can begin to optimize most pages.

Holes in a bucket

With your idle rate calculated, you now have a good, high-level understanding of your page's success – which is 99% more than what most UX professionals have. You're ready to move on to the next level, which is just as simple. Traffic flow.

Think of your webpage as a bucket. The traffic to your webpage is water pouring into the bucket. The bucket has holes in it that you purposely drilled out, of varying bucketsizes. You want the water to flow through the bucket, but in varying amounts through the different holes. The holes are the actions – links and buttons – on your page that guide traffic flow from your page to the next.

Your job is to figure out how big the holes should be. For some pages, like an order submission page, this is simple. Your bucket has one big hole in it – click submit. For most pages though, there may be many actions that can be considered success in varying degrees.

If a page has five links on it, which ones would you hope a user would click on more than the others? Force yourself to differentiate success and figure out which actions are more important. Consider the next 1,000 visitors to your page – what percentage of them should click on which links or buttons? You can define traffic flow by individual link or, if your page has a ton of links, by groups of links. You need to define the percentage of clicks that each link or set of links should receive out of the total links that the main body of your page receives.

Treemaps are great ways to visualize traffic flow. You just make a big rectangle and make treemapsmaller rectangles inside, with each rectangle representing an action or set of actions on your page, and each rectangle's area representing the proportional amount of traffic that each action should receive.

If you really want to get fancy, you could create Sankey diagrams. They're the best visual representation, but they're harder to create.

Measuring idle rate and traffic flow

Google Analytics has a single report – Navigation Summary – that has the data you need. The In-Page Analytics report also has a good view that shows the traffic flow directly overlayed on your page.

Here's your optimization plan

  1. Pick your website's top five pages.
  2. Calculate the idle rate for each of these.
  3. Create a treemap of your desired traffic flow for each page.
  4. Create a treemap of the actual traffic flow for each page.
  5. For each page, print out a screenshot, with the big idle rate number, and the two traffic flow treemaps. Post them prominently for everyone to see.
  6. With new design and content variations, try to lower your idle rates and see if you can get your actual traffic flow treemaps to look more like your target traffic flow treemaps.
  7. Go back to step 2. Repeat ad infinitum, asymptotically approaching perfection.

A word about soft and fuzzy success

Just because you can't measure it, doesn't mean it doesn't matter. Things like brand impression are tough, if not impossible, to measure, but they still matter.

Success measures around engagement can also be tough to get a grip on. If you have an article page in which the primary goal is for the user to read the article, the majority of success is not defined by navigating to another page. Rather it is engaging with the page that the user is currently on. You can optimize measures around engagement, but it is more difficult to do. Start with the easier stuff first – idle rate and traffic flows – and then move on to the trickier stuff.

About me

I'm Mike Padilla and I want to help make the web better. Please help me spread the word below. Thanks!