Business Intelligence in Education – What can we learn from supermarkets?

Part One of a series – and a bit like chapters in a book, chapter one doesn’t tell the whole story, but gets the journey started!

Puzzle piecesFor years we have been collecting data on students. In the beginning, it was data created and collated by individual teachers – in students’ own workbooks and teacher markbooks. And then in early student information systems, we started to collect information on statutory tests, and then increasingly in-school tests. But still, in most schools, data is distributed across lots of different places – information is stored in students’ own workbooks, teachers’ markbooks, teacher spreadsheets, the student information system and in other databases (sometimes held at a teacher or department level, sometimes at school level). Oh, and then there’s the data aggregated and held outside of the school by curriculum authorities, state education departments, and assessment agencies.

But having collected all of that information in different places, and different ways, we haven’t yet reached the stage of using the data to consistently support learning for individual students. Of course, there are some shining examples of where it is being used, but the key word in the sentence is ‘consistently’  – all of the data, connected and used in every school to support learning.

So, the million dollar question is, as John Davitt put it:

  Why does a supermarket know more about my frozen pea buying habits than my children’s school knows about their learning?  

What happens in retail?

Compare the experience of collecting and collating data in supermarkets to the story above. As consumers, we’ve been conditioned by the big supermarkets to share our data, and supermarkets have built central mechanisms to collect and use the data. There’s even a benefit paid for sharing your data. Here’s how it works:

  • Supermarkets sell thousands of products to thousands of consumers, every minute. And that generates a stream of information (What’s selling today? What isn’t? What’s running out of stock? Which shops are making a profit?)
  • To connect that information to individual consumers, they have persuaded us to use a loyalty card every time we shop. Now they can link their generic data to individuals (Who’s buying nappies? If they buy nappies, do they also buy wet wipes? What’s the trend on a consumer’s spending? Are they likely to be shopping at other supermarkets too?)
  • We individually decide to opt-in to sharing our data (by signing up to the scheme, and handing over our membership card every time we shop)
  • In order to encourage us to share our data, the supermarkets have offered us a small incentive – points, air miles, fuel discounts etc

The data that’s collected is potentially massive – and insightful. Everything from what paper you read, to what meat you buy for your BBQ. And by linking that to external data sources, they can go one step further – what products do customers buy when the temperature gets over 30⁰. Where do you live, shop and fuel your car?

Which results in retailers being able to collect and collate enough individual information to help them improve their business, through things such as:

  • Using aggregated data to plan their business – eg Where should the next store be built? What product lines can be expanded? How long should stores  be open for?
  • Using individual data to grow their business – eg What’s the next product line to sell this customer? Which customers can I encourage to shop more frequently? Which customers can I switch to more profitable brands?

And some of this then results in a further incentive for the customers – like offers related to switching brands, or finding that the shop doesn’t run out of your favourite ice cream for the BBQ on a hot weekend (okay, there’s some things that still need working on!)

Why we shouldn’t compare education to supermarkets’ use of data

But is it right to compare what is happening with consumer shopping data to what could be happening in education? In some ways it’s an unfair comparison, as the scale is massively different – supermarkets are dealing with millions of customers, and so they can afford to invest the time and money in building big data models. And there’s a commercial imperative to improve, which results in more revenue and profit – it’s not a fixed budget, so investment in improvements pays back with extra cash.

And there’s also a much more centralised system – for both customer management and data – that results in all of that useful data being seen and used at headquarters, rather than at branch level.

In education, only some data is shared with the ‘HQ’ – state or nationally – for example, statutory test data, like NAPLAN results. Whereas a lot of it is created and stored by the school. Or just by a teacher for their own use.

And finally, the money spent on improvements doesn’t necessarily generate more budget for the school.

The lessons we can learn from supermarkets

Although direct comparisons are unfair, there are some lessons learnt by retailers that might be useful

  1. Individuals are willing to share data if there’s something in it for them.
  2. Reducing friction on sharing data improves everybody’s willingness to share
  3. Everybody in the data chain should receive benefits
  4. Connecting more data sets has an amplified benefit

If you consider how you use data in your school/TAFE/university, are there frustrations about data use that could be overcome by applying one or more of the retail lessons?

In my experience, the ‘benefit’ test is a key omission – eg are you asking teachers to supply data, without ensuring that they receive a benefit for it? (Or worse, do they think that sharing data just gives ‘management’ another stick to beat them with?). And reducing friction is also key – eg if a teacher currently stores their markbook in an Excel spreadsheet, can you read the data from there, instead of making them change?

As I said at the beginning, this is Chapter One. I’m going to come back to this whole subject in a couple of days, but for now, let me leave you with a question to think about (and comment on below?):

What are the other lessons we can learn from retailers’ use of data? Positive and negative?