How to resolve those infuriating analytics discrepancies

Yesterday I was conducting a “Web Analytics Forensics” session with a new client. They posed a common question: The monthly reports on clicks that they were getting from suppliers of their ad buys were off by 10 percent — sometimes even more — compared to their own Google Analytics metrics. The number differences were veering all over the road. Sometimes these vendor reports seemed to overstate traffic, other times the clicks seemed to be understated. When I responded, I was reminded of the reassurances that a friend of mine gives. He’s a pediatrician.

My friend Paul had told me once that the bane of pediatricians everywhere are the late-night calls from parents who are worried about their child’s fever. He doesn’t mind being awakened (well, not much), but he has trouble fully reassuring parents of this fact: Fevers are normal, even healthy. If a child doesn’t run a fever every so often, it’s then that he’d be alarmed!

Don’t take Paul’s word for it. Here is a post in the New York Times a couple of days about on this very subject.

Client concerns about analytics discrepancies are my own profession’s “fever fears.” They can be a distraction from deeper problems. (The doctor in the post mentioned a mom whose child had a fever and abdominal pains. He said his primary concern was the abdominal symptom, but Mom kept steering things back to the fever!)

So the answer to the promise I gave you in the headline is this. For media buy discrepancies, don’t bother trying to resolve them!

Unless you think you are being overcharged for the traffic you’re buying from vendors, in the form of ad clicks or affiliate links, rest assured. I’d only be worried if their numbers were identical to yours. No two systems measure web traffic precisely, or the same way. They are all uniquely flawed.

So instead, focus on what you’re doing with this traffic after it arrives. Are visitors finding what they came for? Are they returning in healthy numbers? And most importantly …

Are they converting?

Usually when I’m called in to consult, the answer to all of these questions appears to be no. Focus your attention, and your boss’s, on these issues. They are the only path to online marketing success.

Finding B2B leads from your web logs

Occasionally I share links to blog posts of others in my industry. Some things are too good to keep to myself. Here’s a perfect example, from Luna Metrics:

We have a customer who considers the SEO we do for her to be one of her “sales channels” and we get ranked along with her other channels. She sends us reports when a lead comes in and when a lead is closed. The other day, I saw that she closed one that was worth not quite half a million dollars. (!! that was my reaction, too.) So I wrote her and said, how awesome. To which she replied,

“Started with google analytics. Saw that they spent some time on the site… sicked Jane on a cold calling mission… after a bunch of calls she found the engineer at the company who was interested in the product. I flew out.. presented… sold and they put out a public bid. Our company is the low bidder and need to send a sample next week for review then release of contract.”

So in case you are wondering, she was talking about the [Google Analytics] Network Locations report, which she mines daily for sales leads.

It’s hard to believe all of this was accomplished by reviewing a typically-overlooked report in a totally free analytics package. Read the rest of their post, and check out this helpful post from the past on how to exploit the fact that many large organizations will self-identify in this report instead of resolving to their ISP’s name.

Visualizing bounce rates using brownie charts

Today on Jason Fall’s Social Media Explorer, I discuss my new favorite data visualization technique — one that I’m starting to move into production with web analytics reports I create for clients. Its official name is the Tree Map, but as I mention in that post, I prefer to call it The Brownie Chart.

That post has an example of how I use brownie charts to show a promising new web metric, the content interest index (CII). My example on the site uses a made-up business, Everything Brownies, with a web address of EB.com.

Note: Yes, I know. That web address resolves to a real encyclopedia site. The reason I didn’t just make up a domain name is you never know when one will go live with a site. I didn’t want to have someone inform me, two months from now, that my blog is now pointing to a porn or gambling site! Unless Encyclopedia Britannica takes a surprisingly sleazy turn, I think I’m safe.

Here is another example of how the tree map / brownie chart can make web analytics reporting easier to understand:

Charting Bounce Rates: “I came, I saw, I puked.”

I agree with Avinash Kaushik that bounce rates are a helpful way to measure how well you’re connecting with site visitors. Actually, he’s a little more enthusiastic than I am, with blog post titles such as this model of understatement: Bounce Rate: The Sexiest Metric Ever? Three years ago, on his own blog, Avinash described bounce rates this way:

So what is this mysterious metric? In a nutshell bounce rate measures the percentage of people who come to your website and leave “instantly.”

They’re the one-page visitors. Yes, they might be finding what they were looking for — but more often than not, these people just didn’t dig the neighborhood.

Avinash has refined his description over time. In his recent, truly outstanding book on measuring web traffic, Web Analytics 2.0, he characterizes bounce rates this way: “I came, I saw, I puked.”

Bounces can be reviewed for all traffic to a site, or only for certain important segments — traffic from search engines is a good example. Reporting of bounce rates can also be broken down by page.

The brownie chart becomes particularly handy for this per-page bounce rate reporting. It helps those responsible assess the severity of a site’s problem pages.

You see, you can’t easily be sure that a page with a high bounce rate really is a problem page. Think of it: If nearly everyone ups and leaves when they arrive at a particular page, but that page gets relatively little traffic, there’s no huge emergency. Content management resources are usually scarce, so it’s better to keep looking, for other pages that attract more page views that happen to have comparatively high bounce rates. It’s those more popular pages that require immediate first aid!

To illustrate, take a look at Everything Brownies’ bounce rates on this brownie chart. The graphic shows all major pages of this fictitious site, and shows the pages as more red if they have the highest bounce rates relative to the others. You should know that size represents the relative numbers of page views. The bigger the “brownie piece,” the more views that page gets.

What does this chart tell us? Quite a bit.

The Holiday Brownie Baking Kit, which I placed my mouse over in this screen capture, has an excellent (i.e., low) bounce rate. It also has a ton of page views.

That means this page is doing quite well in keeping visitors from leaving immediately. Well done! On the other hand, Deluxe Baking Pan is not nearly as successful. Its relative bounce rate is quite high, and because it has the most page views of the entire site, it’s clear this page is majorly dropping the ball!

There are plenty more insights, but you get the picture.

As I mentioned on Jason’s blog, what I like about this charting format is non-math types (such as myself!) can understand these statistics immediately, and know exactly what needs to be investigated further — and in what order of priority. As my friend Bob likes to say, “That’s good stuff!”

I hope you find the potential of this charting technique as exciting at I do.

A Round of Applause for BeGraphic and Sparklines for Excel

This example of a fake report for EB.com, as well as the one on Social Media Explorer, was produced using an “Add-on” for Excel called BeGraphic. The Add-on consists of a whole suite of graphic tools — all based on Excel data and rendered within that application. The particular functions I used were part of Sparklines for Excel within the BeGraphic suite. I urge you to support the folks behind these amazing visualization tools.

How standard email marketing metrics fall short

When you’re trying to optimize the profits of your business, most web metrics are unhelpful at least, and deceptive at worst. But what about the world of email marketing ? Kevin Hillstrom, in his excellent Mine That Data blog, gave this example to illustrate how conventional email metrics look at the wrong things:

Say you have a list of 500,000 e-mail addresses. You send your standard campaign on a Monday. Later in the week, you tabulate your results:

  • 500,000 recipients
  • 20% open rate = 100,000
  • Of the opens, 20% click through to the website = 20,000 visit website
  • Of the clicks, 5% convert and buy something = 1,000 orders
  • Average Order Value = $100
  • Total Demand = 1,000 * $100 = $100,000
  • Demand per Recipient = $100,000 / 500,000 = $0.20 [per customer]

He compares these finding with what you’d get if you did something called a mail/holdhout test. You compare a control group that does not get the email with a test group that does. For instance, he suggests this breakout:

  • Mailed Group = 400,000 Recipients, $300,000 spent = $0.75 per customer
  • Holdout Group = 100,000 Held Out, $45,000 spent = $0.45 per customer
  • Incremental Lift = $0.75 – $0.45 = $0.30 per customer

Much more insightful!

This is why I’m not a fan of open/click/conversion. A mail/holdout test proves the actual value of an e-mail marketing campaign. In this case, we observe $0.30 lift, whereas open/click/conversion yields $0.20 lift. E-mail marketers, why would you not want to know that your campaigns are working 50% better than when measured via opens/click/conversion?

Kevin goes on to provide other interesting observations from his years of doing this sort of testing. You can’t go wrong by following his blog, and trying his approach to data-driven online marketing.