Avinash Kaushik is the digital marketing evangelist for Google and the co-founder of Market Motive Inc. Through his blog, Occam's Razor, and his books, “Web Analytics: An Hour A Day” and “Web Analytics 2.0”, Avinash specializes in how executive teams can leverage digital platforms and data to out-innovate their competitors.

What is the single most important thing that companies should do with analytics?

Don't start with the data. Start by going through the process of creating a “Digital Marketing and Measurement Model [DMMM]”. It is a simple, five-step process that takes you through identifying business objectives, digital goals, key performance indicators, targets (this is key!), and segments.

It puts the management team in a hot seat because they have to help define objectives and goals. Not having these is the root cause of failure for most analytics efforts, because without these two focusing factors, analysts end up doing a lot of data work, but it is not clear what they are doing it for.

The other incredible benefit is that it tells the analysts what to ignore. In a world of unlimited data (for free, if you use Google Analytics or Yahoo! Analytics, etc.), we can keep data puking until the cows come home and still have an infinite amount of work to do. The DMMM says to your analysts: "This is what's important, just these KPIs. Now go and do analysis just for those KPIs and come back with actions we should take — actions written in English — and skip the data puking."

Win-Win.

What’s the biggest mistake you see companies making with analytics?

They obsess about acquisition metrics and outcome metrics and a lot less about behavior metrics.

Acquisition is everything we do to attract traffic. This is important. Outcomes are what happens before the visitor leaves our site. This is important because we have to make money. But it is what happens on your site — behavior — that ultimately determines if any outcomes are going to be delivered!

So I recommend that in addition to worrying about acquisition metrics like Impressions (actually skip that one), Clicks, CPA, and outcome metrics like Conversion Rate and Average Order Value, they also deeply obsess about content analysis by using metrics like Page Depth, Visitor Loyalty, Bounce Rates, Abandonment Rate, and engagement with videos, comparison charts, tools, etc.

Focus equally on acquisition, behavior and outcome metrics and you'll win big.

[Bonus: I have a post on best metrics and a helpful picture at the end. You can read it here: Best Web Metrics / KPIs for a Small, Medium or Large Sized Business]

You talk a lot about silos. What is it that makes companies keep falling into the silo trap, and what practices can you suggest to avoid it?

History. It is just the way companies have always been, and as a company goes from small to medium to large, silos have a way of creeping into the organization. 

From a data perspective, you can incentivize breaking of the silos by ensuring that you measure acquisition, behavior and outcome metrics. Those are typically different teams, but if you declare success only when all three are working, you will create an incentive for those teams to work together. 

On the marketing side, using Multi-Channel Funnels type analysis will incentivize the right behavior because marketers will see that search does not work in a silo, and neither does Facebook nor Email nor Display, etc. The analysis will show that it takes all of those or many of those channels to work together to convert one person. That drives good behavior in a company.

But perhaps the most potent weapon at your disposal is to use the right bonus and promotion criteria. Humans are quite Pavlovian that way. We react to a reward. If you set up siloed rewards, you'll get that behavior. Set up non-siloed rewards tied to the entire company's success and you'll see non-siloed behavior. 

Related Posts