Your Analytics Tool Is Wrong. And It Can Put You Out of Business

That's a fact.

I know this because in my career I've audited the Google Analytics, WebSideStory (now defunct), and Omniture/Adobe Analytics installations of around 50 companies. Give or take. I've personally been responsible for the installation and set up of analytics tools at two publicly traded companies and a few startups.

No matter how big the company. No matter how many people are working on it. **There is always something wrong. Always.

Not one installation I've ever looked at has been perfect.

Note that I'm usually not being called in to audit analytics data. In all cases, I'm being called in to help a company grow, because there is "some reason" the product isn't catching fire. After 13 years of seeing this pattern over and over again, I don't believe this is a coincidence.

When your metrics aren't right, you are flying blind. How do you hope to grow if you can't identify what's really working, and more importantly, why?

Here are three real work examples from just the last month of client work:

Client A, an ecommerce retailer, had a typo in the goals setup of Google Analytics. Because they were using two different checkout pages, they never noticed that they weren't tracking 20% of their conversions. I noticed it when my AdWords campaigns didn't seem to be converting, despite solid click through and near zero bounce rates. After a couple of hours of troubleshooting, and finally blowing up my screen 400%, I found a tiny extra character next to the pipe in a regular expression. Now not only are the AdWords campaigns properly tracking ROI (and showing the great work I did); but their overall site conversion rate just jumped up dramatically which gave the team great confidence in their mission.

Client B, a software company with a subscription component, in a quest to be data driven, installed a different tracking code to the interior/logged in version of their site, so that there was no data on the logged in state to use against life time value metrics. This only surfaced after we began digging into the incredibly high bounce rates on the site. Of course, high bounce rate sites are going to happen when you stop counting the people actually staying on the site...

Client C, a content company, were tracking signups as destination pages, not as the events they've been engineered to be.So they were dramatically OVER stating their sign up rate, and thus making decisions in marketing that could have put the company under. Now that they're tracking events instead of destinations they'll be in a great position to evaluate the cash they want to spend on paid acquisition, and report accurately to their venture capital investors.

I insist on an analytics audit as the first part of every marketing engagement.

How can I help you grow your business if the metrics are wrong?

If your agency isn't helping you with this, let us know. I'd be happy to set up a consultation and an audit that's affordable for nearly every stage of company growth.

In the meantime, a few suggestions:

  • When it comes to analytics, don't confuse precision with accuracy. Just because those charts look awesome doesn't mean they're right. Garbage in = Garbage out. Be as skeptical about the information they contain as you are about your budget until you know exactly what data is coming in and how.
  • Assign one marketer and one developer to OWN marketing data. The marketer should be comfortable enough in code to discuss dynamic variables, script tags, and so on, but doesn't need to actually be able to code. The developer needs to understand javascript and how the site is actually recording critical business events.
  • If you see dramatic shifts in data, assume the code is broken. This is just a practical tip.
  • Data=what. Customers=why. All analytics can only inform you what is happening on your site. If you want to understand WHY people are leaving you have to talk with them. There are tools to scale this for big companies, and tools to do it cheaply for startups, but don't forget to use them.

This post originally appeared on Linkedin.

What have you found that's wrong? How are you keeping on top of this critical task in your marketing team or startup? Let's discuss on Twitter or in the comments.