Just about every e-commerce leader I talk to tells me that they are drowning in data.  Not only do they find it frustrating in that they struggle to find information that is valuable, but just as concerning is the increasing amount of time and resources that they need to allocate to try to make sense of it all.

Analytics in e-commerce is becoming an industry in itself and as we invest more time, money and effort there is a risk that we end up with solutions that make a bad situation worse.

Not all data is valuable all of the time.  Most data is unhelpful out of context. This calls into questions organisation strategies that try and aggregate every data point into a data warehouse, cube or other some such solutions without a filter or expertise that creates an informed narrative.  Whilst ‘knowing everything we know’ might sound the right thing to business leaders, the reality of mass data storage can be one of knowing even less.

Most statisticians will tell you that mass data mining outside of pure science disciplines comes with very large health warnings: how good is your reporting to pull out the right combination of data?  How smart are the algorithms built to pick up patterns that carry meanings?  Can you really be sure its causation, or is it correlation or just pure coincidence?  As soon as human behaviour is involved more often than not you need the understanding, insight and experience of other humans to make judgements.

The problem for central data teams is that they rarely have the in-market and customer context required to judge the criticality of the pattern they can see. So, the impact is often continuing requests for explanation and justification and the consequential distraction of people from the job of growing the business.  ‘Reporting Up’ in the end gets replaced with ‘Justifying Up’ and little or no time or money is saved (especially when you add in the technology costs of creating a single data solution) for very little additional value.

If the Data Cube/Warehouse solution comes with its downsides, the creating of your own dashboard using one of the myriad of self-build solutions might sound a better bet.  Dashboards however come with their own problems, not least the time and effort of making them work.  Regardless of how attractive the visualisation, they too present the same issue as the organisation-wide solutions.  Is anyone looking at it any the wiser about why the data is the data and what, if anything, should you do about it?  So, we are back at the same problem – granting wide access to the dashboard risks merely replace reporting with justification.

And even if you were willing to invest time and resources into a Dashboard solution, there is then the real problem associated with e-commerce data – believability. You’d be amazed by how many web analytics platforms we’ve seen that are poorly set up: some so much so that they deliver erroneous data that can drive entirely the wrong action.  At the same time, much of the data that is presented to us by agencies or marketing platforms is at best uninformative and at worst self-serving.

The common problem here is that there are very few proper scientists working in e-commerce teams or digital agencies.  Like it or not, GCSE maths just doesn’t cut it as far as understanding the nuances of data and even social science degrees such as business and economics don’t often build the data analysis capability that can turn data into cash.  The result is that despite the increasing amount of time and for many businesses the growing number of people employed in e-commerce, their teams are not delivering the performance expected by their senior leaders.

So how can you save time and resources by not reporting or justifying merely to service your corporate machine?

The answer lies in shifting from thinking about data, to understanding the story behind the data.

As an e-commerce leader you want the data you need to ‘understand the story’ of performance.  That means getting a view of the totality of performance and doing so from no more than about 40 or so data points.  From there you can dig deeper and further if required.

What would make this invaluable is if that these data points don’t just arrive through well-presented visuals, but that they are accompanied by a thoughtful and insightful narrative written by a scientist where you are confident that you are getting the unvarnished truth.

There are two ways to achieve this:

Do it yourself:

• First, build a model of your e-commerce system and understand how it works
• Identify the key metrics across the system that align to your strategy and goals
• Invest in a pure science team who can quality assure the data, do the analysis and write it up in such a way that you can take actions and impact the business

Or:

• Explore an off-shelf solution that has already developed a model of the e-commerce system, mapped all the available metrics and uses the science team relied on by the e-commerce teams in brands such as boohoo, WH Smith, Pets at Home, QVC, Heinz and Puma to tell the story behind the data.

You’ll find more on this approach here – and, by the way, for one domain reported monthly, as well as saving you hours and hours, it will also cost you far less than you think.