Beyond Web Analytics: A Holistic Approach

I'm a Web Analytics Consultant working at a 'google oriented agency' which means I work with Google Analytics and the entire Google Marketing Suite. This has a lot of advantages of which easy one-click-integration between their advertising & analytics products is one of the biggest. (oh, and it's 'free').

But, just like any mass-produced product, there are limitations to the abilities of Google Analytics. When it comes to pure web analytics I still think it is the best software out there, and on the mobile analytics front its definitely improving fast. But as we keep preaching about 'breaking down the silos' it makes less and less sense to have a website analytics tool.

Other Datasources

Think about the systems in your company. You've probably got some kind of CRM to keep track of clients and sales-efforts. There might be an ERP which keeps track of products, prices and inventory. B2C Companies most likely have a customer service tool and maybe even a callcenter; the list is endless.

In my opinion, a customer and the information relating to this customer should be tracked across all these services. Simply knowing what Google AdWords campaign a user came from when he filled out a lead generation form isn't going to cut it. How was this lead followed up in the CRM, did it become a client, how much money did he spend, over what period of time? These are all questions that are relevant when evaluating the effectiveness of the Google AdWords campaign.

And it's not just about being able to correctly evaluate your marketing efforts (although it would make my current job a lot nicer), it's simply about making truly well informed business decisions regarding the cost of customer acquisition and the customer lifetime value.

But isn't that Business Intelligence?

Way before web analytics, there simply was business intelligence. The problem I have with BI is that the only solid implementations I've seen of it is at big corporations. And even then it's usually focussed at internal processes and based upon ERP & Accounting software.

Over the last ±15 years web analytics has seen a tremendous growth, and companies digital outlets (websites, mobile apps, social media) have become their biggest source of customers. Thanks to software as a service business models we have been able to give incredible insights regarding web analytics to small- and medium businesses instead of just big corporations that could pay for proprietary software. This has made us all hooked on data, and now we want more!

Data Warehousing

So what should be the next step? In my opinion every business should have data collection as a pillar in their businessmodel. If this is the case, you quickly figure out that simply storing all your data at Google isn't a viable option for the future. Sure, the data can ALSO be at google, but you'll at least want your own copy! And besides that, Google Analytics only focusses on your website data, what about the rest?

Luckily, companies like Amazon Web Services have made it incredibly easy to set up your own "Data warehouse" with minimal expenses. So you don't have to invest in a server installation and a systems administrator to monitor it! But how are you going to capture all this data from all these systems in a way that makes sense?

Event Driven Analytics

Event Driven Analytics is a term that is popping up more often. These "Events" we're talking about include your standard web analytics events like "pageviews" but they might also include events from your callcenter or CRM system. Basically, everything that uses computer technology should be able to send "events" to this event driven analytics system.

This creates a "shitload" (official technical term) of data that is being fired towards your datacenter. Simply piling it all up in a good old fashioned database table won't suffice if we want to analyse the data and get insights from it later. Especially not if we're talking about millions of events a day, or "Big Data" if you would like.

Snowplow Analytics is a great example of a young company that is tackling this problem. They are building a system that, through "data schemas", should be able to handle a wide array of input and store it in a way that is usable for businesses & analysts. Im not going to dive into technical details, but if you're interested I suggest you read their documentation.

The Implications of Data Warehousing & Event Driven Analytics

The upside of using your own data warehouse and a event driven analytics platform like Snowplow is that you have complete control over every aspect of your data collection & analysis. This is nice from a data-ownership perspective as well as the flexibility it gives you to integrate new tools and collect data from them.

The downside is that simply pasting a piece of code somewhere and waiting for the data to pour in is no longer working. You will need to employ or hire people who think about what to track, how to track it and how to make the data accessible to the rest of the organisation. Besides that you will need someone to keep an eye on the data that is being collected and the effectiveness of that process (you don't want a messy huge slow data warehouse!).

This is exactly why "Database Architect" and "Data Scientist" and related job descriptions are one of the most highly recruited ones on LinkedIn right now. This, unfortunately will make it a bit more expensive than your free version of Google Analytics and employing one fresh-out-of-college web analyst to produce you some nice reports. However, I feel that the insights gained by implementing this truly holistic way of analytics will eventually outweigh the costs of implementing and maintaining it.

Hopefully this post got you thinking about analytics at your organisation and what the future could possibly bring. Please ask any questions or remarks you've got in the comments section below and I'll be sure to answer them to my best abilities!