Category Archives: Analytics

Analytics – Details Matter

Recently did an analysis of one our businesses – Starlit Citadel – and noticed that conversion rates had taken a drastic turn down.  It went from 1.56% to 1.15%.  That’s a horrid, horrid drop and even though we had a lot of competition show up recently, it seemed like a truly significant drop.

Now there’s a few things we could do.  Drop our prices and kill our margins, which if it had allowed our conversion rates to stay the same would have resulted in overall higher sales and a marginal improvement in profits (before taking into account additional staffing costs, etc.).

On the other hand, perhaps we should look more closely first.  So, here’s what we looked at to figure out the change:

  • Compared it to previous year stats – nope, still a drop for sure
  • Compared it against traffic sources – all over drop pretty much. Not good…
  • Compared it across new and old visitors – hmmm… new visitors were the one who killed it most
  • Compared it based on location – ah, this is more interesting.  Traffic had gone up significantly in countries outside of Canada.  This pushed down our conversion rates because people who aren’t Canadian just aren’t that interested in us.

As we said, details matter.  If we had decided to take steps to decrease our prices and increase our conversion rate, it would have helped but only a bit.  Instead, we’d have suffered more of a loss because the traffic from outside of Canada just don’t care about us. A price drop just wouldn’t help.

So, details matter.  Dig, dig, dig to understand what happened and is happening.

5 Analytic Key Performance Indiactors

Let’s talk about analytics.  If you own an e-commerce store, you want to know what to track.  So, let’s talk about the key perfomance indicators you’ll want to watch.  Remember, if your focus is a specific geographic region, you’ll want to track both for that region and overall with particular emphasis on your target region.

1. Sessions

Sessions or visitors if your analytics program uses visitors instead.  This is the top of the line metric, and will give you an idea of how many people are ‘coming to your store’ each day.  It’s always good to move this number up, especially if you are sorting by your target market already.

2. New Users %

Here’s one that most people don’t think of.  As a rule of thumb, you want to get around 50% new users coming to your site on a regular basis.  Any fewer and it means you aren’t generating enough new leads to sustain and grow your business.  Any fewer and your site is not ‘sticky’ enough. Of course, new sites will find their New User % really high, but you do want people to come back and you need to move those returning customers % up.

3. E-Commerce Conversion Rate

Industry average is 2%.  Frankly, I’ve never been in an industry where I’ve managed to get 2%.  Now, this might mean I’m just really bad at it – but it can also be an artefact of the industries I work in.  Realise that an overall benchmark like 2% can only give you the roughest of numbers to begin with.  You should however be able to hit between 1- 2% for the most part.

Most importantly, what you do want to do is adjust your site and site design to push your e-commerce conversion rate up.  Whatever you do, you want it to go up – not down.

4. Average Order Value

With this, sessions and e-commerce conversion rate you can work out what your revenue is.  More importantly, you know what your average order amount is and what you need to do adjust the value up.  In general, you want your average order value to go up (unless you figure out a way to increase volume transactions instead).  If this number goes down though, then your product mix needs fixing.

5.  Average Pages Viewed

Lastly, average pages viewed shows how engaging your site is.  Bounce rate will tell you if a customer will stay pass the first page, but it’s average pages viewed that tells you how far they have gone in the site.  The deeper they go into your site, the more likely they are to buy from you and/or add more products to their shopping cart.

I find this number better than time on site because if you improve your site’s speed (performance); you’ll find that the time on site will drop but average pages viewed should stay the same (or more likely, go up).



One of the most interesting articles I read a while ago discussed how a website owner liked to use two analytic programs on their site.  The two discussed was Google Analytics and Pikwik.


The reasoning was really simple – analytic programs have a tendency to update and change the way they track data without informing you.  As such, it’s often useful to keep track of the data using 2 different programs so that a drastic change can be tracked backwards based on previous data.  Or put another way – you can tell if it’s your site or just the software.

There’s another subtler reason buried there too of course – Analytic programs track the same information differently.  It can often be useful to have 2 programs track your data separately if you understand how the programs track / categorise that information extract more knowledge.

Sometimes, the reasoning can be even easier – sometimes, one software just presents the data better.  I definitely enjoy the presentation by Pikwik where it automatically tracks sales by categories for us, something Google Analytics doesn’t do without a lot more work.

Hummingbird & Search Query Unknown

So, Google’s been altering the way it works as usual.  Let’s talk:


So, Hummingbird just released (okay, it was announced released – it’s been around for a month).  For our sites, we’ve seen a small drop off in traffic – not enough to say it had to do with Hummingbird rather than general movement.  However, interestingly enough our page views have gone up which shows better engagement so it might actually be Hummingbird doing a better job.  Conversions took a drop but not significantly, which could be attributable to any one of a milllion things too.

Overall, Hummingbird seems to be a non-starter since it targets longer tail words anyway; and those have always been harder to focus on when you are an e-commerce business like ours.  With so many new products appearing all the time, it’s just tough to generate any good in-bound links to a product.  Worst, more often than not those links just disappear anyway.

Search Query Unknown

What has been worst for us is the complete rollout of the (unknown) search query use by Google.  That hurts – a lot.  Sure, other non-online marketers have been dealing with the lack of data forever, but we’re online – we get data damn it.  It really is frustrating, and will make SEO so much harder in the future.  Not knowing what the number of search queries for a term and whether it’s worth going for makes it more a blind guess.

Mind you, Adwords is doable and combining search rank data with page data will help; but that’s a lot more work.  Doable work, but more work.  Joy…


Keyword – Not Provided

Last year, Google decided to start creating ‘private’ searches; basically automatically removing search result information from all logged in users.  At first, they stated that it’d only affect about 2% of your visitors – a small number.

Utter bullshit.

These days, they take up 28% of our visitors from Google Search.

Why does this matter? Well, more information is better.  But worst, it made Y-o-Y comparisons much more difficult.  After all, 28% of all visitors are being ‘wiped’ from our results – but there’s no way of knowing which results they were wiped from, so it could be 100% of these visitors from a search results to 5% or something.

It’s worth keeping an eye on and better yet, keeping an eye on the rate that this increases.

And if you want a workaround (work being the emphasis here); keyword data is still being given to Advertisers; so log in to your Google Adwords account.