Problems with Big Data in Grocery
It is clear that Big Data is hot these days, so hot that it may even be passé - generally a good sign that a trend has made it main stream.
There hasn't been a lot of good studies on the implications of this trend for the grocery industry, but the infographic below comes from one them. The study, titled Moving Forward with Big Data: The future of retail analytics was put together and can be downloaded from the team at bricksmeetclicks.com.
The infographic below shows some of the highlights from the study (you can click the image for a high resolution version). We thought it worthwhile posting as there were some interesting talking points.
Although most respondents of the study indentified the need or desire to leverage Big Data tools, and even 65% of respondents said they were involved in a Big Data project, what really intersted us were the perceived barriers to entry.
The top three concerns with implementing Big Data strategies were:
-
Concern about not using data already available (58%)
-
Lack of capability to implement insights (57%)
-
Not enough specialized expertise
These three concerns are top of the list not so much because they are necessarily real justified concerns, but more so because they are a response to the type of Big Data solutions that grocery retailers are being pitched or sold. Grocery retailing is particular, with particular needs, and particular data sets and value pools. Big Data solutions that work in other industries do not directly relate.
With the aim of exploring this further, I will detail some examples of why we believe that these top three responses are more a response to the Big Data solutions in the market, rather than Big Data as a strategy.
Concern about not using data already available:
This concern comes not only from the fact that grocery retailers are not using existing data sets to generate insights, but that they are getting pitched to use Big Data strategies on new data sets that are irrelevant for their needs. The thing with Big Data strategies is that it's about extrapolating some kind of insight from vast amounts of data that don't necessarily have a lot of value individually or in small volumes. The problem is that grocery doesn't always have the volume of relevant data needed - take sentiment analysis from social data.
We have seen grocery retailers pitched (and sold) on leveraging Big Data strategies on their social data to measure sentiment of their customers. The problem with this is that the large majority of grocery retailers have very little social data to mine. In the very brief table below i've researched the Twitter mentions for 4 random retailers (below) and you can see that with the exception of Walmart most retailers don't have anything close to a deep enough data set to be reflective of their customer set.
Twitter mentions:
Walmart - 8mm mentions, Meijer - 208k mentions, Schnucks - 22k mentions, Shoprite - 87k mentions
Lack of capability to implement insights:
Many of the big data solutions that we've seen in the grocery industry promise a bevy of insight on customer behavior, sentiment analysis, inventory analysis, marketing trends; but very few of them tie directly in to practical ways to implement the insights. A large part of the reason for this is because it's not easy to implement these insights in grocery due to the complexity of the path to purchase (or more importantly the multiple paths to purchase).
By definition, the path of purchase to BestBuy to purchase a single item is less complex than the multiple decisions being made at each grocery trip. Grocery Retailers need simple insights with direct practical and flexible applications; such as being able to customize shelf tags, customizing direct emails, delivering custom content online or via mobile.
There are few Big Data solutions in the grocery market that are tied to a holistic path of implmentation and this is another reason why the industry is weary of the "Big Data" tag.
Not enough specialized expertise:
At face value this is an understandable concern related to implementing Big Data solutions in grocery, but again, upon closer inspection we suspect that this too is a reaction to the "black box" magic of Big Data as it is often sold in the market.
The number of times we have seen Big Data solutions sold as a "Black Box" Solution - plug this in to our black box and watch as the result comes out the other side - is too many to count. It seems to be a basic axiom of Big Data strategies - and a great promise at that. The problem is that if anyone is going to take the risk of championing a Big Data solution, they need to be able to have some basic logical understanding of the process before they can put their support behind it.
We would argue that a reason for concern about not having enough specialized expertise is the uncertainty associated with "Black Box" solutions.
In truth there can be a need for some specialized guidance, but it is more important for grocers to be firm on the questions THEY want answered and they should instead demand that the providers of Big Data solutions gain specialized expertise in their concerns.
And, it is with this last point that we will sign off on from this post.
If there is one thing we have learnt in all our dealings with data and our clients, it is that THE single most important thing is to ensure that the clients QUESTIONS are clearly outlined and detailed before any work can proceed.
Insight from data analysis is only ever as powerful as the questions that are asked - this is where competitive advantage will be found in the future.
Big Data in grocery | in
Big Data,
LabelINSIGHT,
grocery 



Our mission is to make vast amounts of food label data accessible to brands, developers as well as government entities, while keeping it simple to analyze.” - Anton Xavier, CEO 









