Jason Finch explains why the quest for masses of perfect data may well be futile for now.
I’m not defeatist by nature but the chance of manually finding nuggets of gold in vast amounts of ludicrously detailed data is small. Unless data is absolutely perfect, one can develop tunnel vision searching for facts that can never be complete. Very few companies have perfect data.
Analysts can delve deeper and deeper, slice the data in more and more ways, bring in numerous data streams, and end up merely being blinded by the results.
For someone who has been evangelical about the use of connected web technology for more than 20 years, I can be somewhat technosceptic; and the fact I’ve built businesses that profit from exploiting vast quantities of online data hasn’t prevented me thinking that so-called ‘big data’ is a concept that consultancies selling data analysis have bigged-up fantastically.
Go back to basics: Get your data right before you try to go ‘big’
If you have enormous budgets and access to talented statistical analysis gurus who can sit all day modelling customer habits with Markov chains and poring over multivariate test results, dive in and see what gold they can pan.
However, as wholesalers launch new online ordering channels, there are some far more straightforward, data-related issues to address rather than worrying about how to analyse real-time information that’s being generated faster than ever before.
You may not be collecting the right data from your seedling web operations and you may not be doing it in a way that enables you to blend it with your existing data-gathering effectively. There are some fundamental concepts that you need to nail before you go looking for nuggets of data gold and you can learn from those who went before you.
Living in a multichannel world: Learn from what retailers have already done
Although the online world collided with the wholesale sector a year or two ago, I’ve noticed buzzwords rife in retail ecommerce have not yet appeared: phrases such as ‘multichannel wholesale’ are not common.
I wonder if that is because wholesalers have always offered multiple ways of ordering: telesales, box ticking in printed brochures given to reps, faxed order forms, electronic point-of-sale (EPoS) and so on. To most wholesalers, adding an online ordering channel is not as novel as it was for high street shops.
Then I wondered why wholesalers took so long to discover and implement web systems and why so many are still using it only for fairly basic order capture, rather than targeted upselling and other smart things.
Well, retailers had to develop brand new processes so they looked to the emerging technology to answer their prayers. This enabled them to come up with amazing solutions and novel processes based around what the new technology could offer.
Wholesalers, on the other hand, already had the multichannel theories and processes and so many have used those old models – fit for faxes and ticked boxes and order capture – even if it meant clunky technology workarounds that convert online databases into crude paper systems or Excel spreadsheets, and vice versa.
But whatever the reasons and history, retailers were able to start the more data-intensive, online journey from a different perspective and I believe it’s one that wholesalers need to start following as their businesses evolve.
Everything is coming together: Develop a single view of each of your customers
A newer retail term is ‘convergent commerce’, referring to the fact that the many ways in which a customer interacts with a business are coming together; they are blurring, no longer so distinct.
Retailers used to consider old and new channels rather separately, but their customers increasingly choose to blend the ways they interact, so can no longer be labelled solely as ‘an online ordering customer’, an ‘in-store customer’ or a ‘direct mail customer’.
Many wholesalers appear to still treat channels quite discretely. Different channels often have different people responsible for their sales and marketing and for analysing their data. This has to change.
While customers who predominantly order through reps or by fax may top up using telesales, as they move towards ordering online, you’ll find they’ll expect far more flexibility in ordering and customer service.
You need to develop a single view of each customer that completely blends their interactions with you into one core set of data – one that is always extremely up to date, is actually complete and ultimately will be housed on a single modern platform.
Invest in a new foundation: Don’t try to reuse old technology
Don’t try to shoehorn the far more flexible and advanced concepts available within today’s connected real-time technologies into old systems that were mainly designed to cope with more antiquated ways of interacting with your customers.
Bolt-ons to systems and processes that are a little rusty under the bonnet don’t work long-term in the new connected world. Retailers have discovered this: many of their systems are being recreated on a foundation of new, web-based technology.
For example, the c-store EPoS of the near future will merely be yet another thing that talks to different suppliers’ centralised systems in real-time, permanently connected and communicating through a standard web language – that’s how Google, Facebook, YouTube, Twitter and other online services enable third-party applications to interact with their data.
With an increasing number of wholesalers launching online ordering, the many retailers trying it out will start to expect (or, even worse, demand) features that a traditional disconnected EPoS system simply cannot provide. You may need to deal with an increasing number of customers doing significant re-ordering split more evenly across three or four different channels. You cannot split the responsibility for understanding the data from those across different people.
It may seem an uphill struggle, but at some point you will need to bite the bullet and create a data strategy that focuses all your disparate data feeds into one place. This one place should translate data from various systems into a common format and store it in such a way that any other system can easily request and use any bit of information it needs.
You need to build one authoritative repository for all information. Once you’ve done that, then you can start work on ensuring every piece of data is absolutely perfect.
