Big data sucks!
Updated: 07 Apr 2015
Ok, Big Data doesn't have to suck, but far too often it does! Why? Read on...
Over the years I've been evangelising the power of data and why measurement it important - in short, it's all about the execution.
And this is where my frustration with 'Big Data' - and other vendor pushed terms - stems from. I've been working in the analytics space for around 11 years now and in digital since the mid-nineties. During this time there have been lots of buzzwords and there has been a lot of change in the market place: first web analytics ("Where it's at" - according to M.C. Hammer), then Real-Time, and now Big Data.
Don't get me wrong, the concepts are sound, it's just that more often than not, these buzz words are used without any thought as to why they are being used and what impact they will actually have to the business. As such companies spend a large amount of their capex budgets on trying to implement a solution without actually ensuring it will deliver what they actually need - the ability to effect change - or how they will maintain it.
A lot of the projects I work on stem from a client saying - "I just don't trust the numbers".
In my last roles, as Director of Measurement at Seren, I was frequently asked to help our clients move from one technology vendor to another - we were always technology agnostic - and the issues were always the same. Our client had spent a lot of money with a vendor X over the last few years and they haven't delivered the promised value and so the client was therefore moving to vendor Y.
So, why pick on 'Big Data'? Well, simply because it's the current buzzword. The concept of Big Data, where your various data sources are pulled together, is a very powerful one and it really can bring significant insight and value. So what's the problem? Well, there's a big But... it will only happen if your organisation is geared up for it.
So, to help with this... Meet D.A.V.E. - If you can answer 'Yes' to all of the following, then you might be in good shape - if not then think again:
- Data Access: Do you know what data are available? Who 'owns' it?
This isn't rocket science but basic logistics (yes, boring I know!). This point often comes as a surprise to business owners, but data are often stored in old proprietary systems that no-one knows how to access properly. I've worked with very large brands who are still using a computer under someone's desk to run critical databases - everyone knows that this is loonesy, but it happens all the time. I"d therefore recommend audting your measurement ecosystem to help document what measurement systems are in place.
- Accuracy: Do you trust your data?
Frequently there is a significant gap between how data are captured and how the data are distributed. Understanding this gap is critical to truly understanding the data. Data capture is a complex business and requires a significant investment to get right: both during implementation and on an on-going basis. I actually prefer using 'Meaningful' or 'Consistent' data as the term 'Accurate data' implies that it is a true reflection of reality - it never is. Often business invest too much time trying to get numbers to add up, rather than focusing on whether the data are consistently captured and can therefore give a meaningful reflection of reality. I'd therefore recommend completing a data audits followed by a Gap Analysis (i.e. the difference between what you want to achieve and what your able to achieve) projects to help businesses focus their efforts on the most valuable data.
- Verbalise: Do you know what does 'good' look like?
This is the biggie. By far and away the single biggest challenge for most businesses I've worked with, was their ability to effectively communicate. This often resulted in a disconnect between executive business owners and the technology facilitators. Communication is about getting your point across, about telling a story that engages with your audience - sending a report with lots of numbers is NOT communicating, it's a mess and it's lazy! As such, consider producing a measurement framework which details: what; how; and why; you should measure. The framework should cover: people, i.e. the skills available; process, i.e. ability to share and action insight; and technology, i.e. the tools available. Secondly, consider how to practically get the measurement framework up and running by producing an Implementation Roadmap. This will contain a detailed and prioritised timeline of recommendations designed to reduce the gap between what your business wants to do and what they are currently able to do. This includes upfront 'quick-wins' to get the ball rolling along with both mid- and long- term actions to improve the available skills, processes, and technology stack.
- Execution: Can you actually do anything with your newly gleamed insight? After all, "it's all about the action, baby"!
The final critical part - can you actually implement the recommendations? I was speaking with a client a few days ago, where a project had been significantly delayed due to their analytics platform trying to load data files in near real-time - their goal was to have real-time data. However, actually implementing changes to their web site typically takes at least one to two months and they are now struggling to keep executive buy-in for the whole process.
'Yes' to all of the above points? - then you're doing great and we should talk about how you make full use of your rich and insightful Big Data :)
'No' to a few or all of the above? Then people like me are here to help change that!
Either way I'd love to hear from you!
Be the first to comment! Add a new comment below...
Add new comment (subject to moderation):