How would you count a pageview?


In my day job, I work, among other things, with numbers and stats. That makes me interested in the stats space at large and I often watercooler-discuss what counts as a pageview with my colleagues. Which isn’t quite as simple as it might sound at first. Disclaimer: the following is merely the product of my own observations and musings. No opinions presented here represent those of the company I work for.

So, you type in an address and wait for the page to load. That’s one pageview. Right?

Well, what if you were browsing Google with the new Chrome prefetching technology; your browser might load the page in full without you ever seeing it. Is that a pageview? How about unique visitors… what if you reload a page, is that one or two pageviews? What about an AJAX powered photo gallery that loads inline, does each photo count as a pageview? And what about infinitely scrolling pages? Do they count as a single pageview or as an arbitrary number of views based on the length scrolled or time on the site?

As you can imagine, there are numerous ways to interpret these mechanisms and some boost pageviews quite a bit. Considering ad-revenue or measurements of overall service popularity is sometimes tied to number of pageviews, stats are bound to be all over the map depending on who you ask.

I don’t presume to have the end-all be-all answer to how pageviews should be counted, there are legitimate challenges here. As the web evolves beyond pages and hyperlinks, the way to count statistics is bound to change. Perhaps in the future it will make more sense to count unique views correlated with time on site as a measure of success? I don’t know, but it seems to me it would be sensible for there to be an independant standard for pageview-count best practices. Like with organic food, you could slap a sticker on your stats package and proclaim the principles with which you count pageviews. I can see this being in advertisers interest as well; I mean why not put your money where the pageviews give the most bang for your buck?

Abstract helper monkey paintings for Stats

Not many days ago, sent out annual recap emails to a number of blogs, including TechCrunch. Part of that email was an “Attractions in 2010” list of 5 top posts of the past year. This section included an image, chosen from one of those top posts. Not all blogs we emailed had such an image, however, and for those that didn’t, I painted a number of “abstract helper monkey paintings inspired by the stats”. Here they are:

abstract-stats-1 abstract-stats-2 abstract-stats-3 abstract-stats-4 abstract-stats-5 abstract-stats-6 abstract-stats-7

Overhauled Stats

It’s something I’m completely proud to have been working on for the past few weeks (along with the incredible Automattic team!): has a new overhauled stats page. This first round of the overhaul brings sortable dashboard-esque boxes to the stats page, and a revamped non-Flash, iPad compatible histogram bar chart.

On the whole, the community seems to like the new look, but a number of people miss the old line-chart and a few others find the new design boxy and uninspired. Which is their prerogative and I won’t spend too much time defending it, other than say that a lot of thought been poured into the revamp, and personally I think and hope it’ll grow on people. As for histogram charts, there are a lot of good, semantic, reasons for going with them. But my personal favourite is the large hit area each bar provides, for when you want to look at individual day stats. Penny for your thoughts?

Embarking on a Redesign


Sometimes, a restructuring is necessary. Throw out all the old, and bring in the new.

When To Redesign

There is no finite formulae that when used to crunch numbers while wearing white lab coats can tell you exactly when you need to redesign.

On one side there’s the hot-air department who says: “Redesign every sixth month! Next time we need taxonomic navigation!”. On the other side there are the designs that just work and have done so for a long time.

Yet, when combining the idea of a refresh with the experiences of a tried and proven design, we might end up with a third option. No, not a re-align, a redesign based on purpose, past experiences and the need to move forward. What should we call this? I’d like to call it a redesign done well.

In the end, there’s nothing wrong with moving forwards, and so it essentially comes down to two things: do you want to, and can you make it better.

Before embarking on a redesign, it is a good idea to deal with these two issues, before you lift a pen.

Let’s say you want to redesign, and you think you can make it better. Let’s first note that making it better rarely means improving the visuals. After all, we’re doing the christmas-tree here, not the decoration. Still, a christmas-tree is in itself a proven design—can we improve this?

Focus! Identify Goals To Achieve

If you’re one of the lucky ones, you signed up for Google Analytics when they were still accepting signups and you later on read Mr. Cabanillas review of Analytics Overlay. If this is the case, take a look, and see where people click:


Ahh. Where people click. As you’ll notice, most of the links shown here are actually being visited. Some are visited more than others. Particularly orphaned sections seem to be Installments, Photography, Wallpapers and Noteworthy entries / illustrations.

One could say this is understandable since those sections are more static and less organic than the journal you’re reading now. Even so, I would like to drive more clicks to those sections. As such, this in itself could be identified as a goal, whether this means improving usability or something else.

You’ll notice that the Archives and Colophon sections are fairly equally visited. Those are sections that serve very specific purposes and there’s no sense in driving more traffic to them. Maybe they don’t deserve to be primary navigation?

Ah, two goals already. Now how about some numbers; Analytics tells me that I get approx. 2 pageviews per every single user. Whether that means the RSS feed and a single article of interest, or if it’s the journal and the comments for a single entry, it’s ever so slightly disappointing. Showing more points of interest may be a goal, or maybe it’s just goal #1 restated?

Three goals for improving the end-user experience. But what about my personal experience? Making it easier for myself always increases my productivity.

  • The practical issues of uploading an installment should be easier for me.
    This probably means automating various aspects of the process.
  • Writing and interlinking should be easier.
    Lots of things to do here, but some of the actionpoints could be improving the post display, readability, page navigation and search.
  • Monthly rebranding should be possible.
    A while back, I asked my visitors whether a monthly rebranding was a nuisance to them or not. Essentially it wasn’t, as long as the main branding of the site shone through.
  • I want more space and less scrolling.
    Does that mean liquid width design?
  • Make it possible to move good content bubbles to the top, even if they’re old bubbles.

Name It!

Add value to your work. Give it a name that sums up the core values of your redesign. If a new idea pops into your head during the design phase, hold it up to its name and see if it fits.

Seeing as I need to unite sections of the whole that have drifted apart, I call it Pangea;

Pangaea (Greek for “all lands”) is the supercontinent that existed during the Mesozoic era, before the process of plate tectonics separated the component continents.

Google Analytics, Observations

I’m using Google Analytics, the new Google website statistics tracking package.

So far I have limited stats available, but even in this phase I’m noticing a few things worth mentioning.

  • The primary target audience is execs wanting to track the effect of their online marketing efforts. This is reflected, not only in the design, but in the wording and setup process for the tool; ROI, Conversion Summary, and “View as Executive”. This is a good idea, as in my experience execs are the first to ask for just that.
  • It’s going to be a huge success for Google. It’s a simple, free service that provides in a simple way, what Webalizer has failed to provide for years.
  • Analytics is using a visual identity which is distinctly different from Googles other doings in a number of ways.
    • For one, Analytics is using TradeGothic for the logotype. Until now, Google has been using a bold, colored Arial right below the Google logo itself. See Google Image Search, for instance.
    • The orange top-bar colour is much sharper than the pastel color scheme Google usually sticks to.
    • The orange top-bar is bordered all the way around, instead of only a topmost border like Google Local (which is a pretty traditional Google design).
    • Icons have their own distinct style (Print, Export to Excel, etc.), as compared to other Google services.
    • The login information bar in the top-right corner has it’s very own style, as compared to Google Personalized.
    • The navigation bar to the left has a unique design which I haven’t yet seen anywhere else. Gmail is probably the Google product that has the most UI, but even here there’s not much in common with Analytics. Even the drop-down arrows (Analytics navigation / Gmail labels) are different.
  • The main Google Analytics site is also sporting the new design. Notice the stark orange ribbon? Notice the dropshadow inside? Also take note of the dashed lines, the “gray-as-opposed-to-black” text color and the use of Verdana.

So what does this mean? Well, as I see it, it can mean two things.

On one hand, it might just be a mistake. I’ve seen many Google products with little visual conformity. In fact most standalone Google applications have their own skin; Google Earth, Google Talk, Google Desktop, Google Picasa. Four different products, four different interfaces with only the logo in common (and in the case of Google Desktop, not even the logo in common). If they had only stuck to plain UI Widgets (the standard OS controls) instead of styling the lot, they’d atleast have that in common.

On the other hand, and this is what I think is the case, it can mean a more specific design for a specific target audience. I’m sure there are lots of execs at Google, all sporting the same bullshit bingo, yelling “Synergy!” from time to time. This breed craves stats. Detailed stats, that are simple and easy to understand, but draped to look complex, meaningful and slightly intimidating. Would the friendly Google design work here? No siree, time is money and and there’s no time for simplicity here!

What I would like to see, should Google want to stick around for a while, is a clear and distinct visual direction for all their products. Make a visual identity, and stick to it. This doesn’t negate the possibility of targetting specific audiences, it only means there’s going to be a visual connection between services—besides the logo. Four different skins are three or four skins too many. Four different designs are three designs too many.

I’m convinced it’ll benefit them in the long run to plot a course for their visual efforts. For now, we’ll have to settle for the contents.