Online retailing means the industrialisation of analytics

Figleaves founder and eCommera chief scientist Michael Ross argues that online retailing requires a new type of action-oriented business intelligence

Online retailing requires a new type of action-oriented business intelligence, according to Michael Ross, eCommera chief scientist and founder of online lingerie retailer Figleaves.

The former McKinsey consultant and Cambridge maths graduate contends that e-commerce will thrive to the extent that it industrialises the knowledge work performed by data scientists, or analysts.

“It is similar to 200 years ago, at the time of the Industrial Revolution, when craftsmen were replaced by factory workers. At that time some people would have thought: ‘Why would you buy something made in a factory rather than by a craftsman?” he says.

Ross (pictured) sees Jeff Bezos, the founder of Amazon, as the trailblazer for a business philosophy which sees e-commerce as owing more to the continuous industrial process improvements of Toyota than Tesco.

“If you compare e-commerce to physical retail, with [the latter] you can run a business with a small amount of data. Let's say you have 100 stores - you can look at your inventory by category and get a pretty good idea of what's going on in your business,” he says.

“And while there are huge opportunities for doing data mining, as Tesco does, that's not a prerequisite for success. The economic structure of physical retail makes it fairly simple to work out what's going on.

“In the online world, a few things change. You suddenly have a whole raft of daily optimisation decisions around marketing, merchandising, and so on. You've got a lot more decisions to make, and a lot more data with which to make them.

“There is an interdependency online between marketing, merchandising, the site managers and so on. The decision paradigm is much more complex.

“For example in physical retail, either a product sells or it doesn't. In the online world if, say, your jeans are not selling, is it a viewing issue or a conversion issue on the website? The bottom line is this: if a product is not selling online it could be an issue for 10 different people”.

Read more on business intelligence in e-commerce

Gathering retail intelligence

SAP buys e-commerce company Hybris to move into CRM market

Trends and tips for using business intelligence and analytics in retail

Mobile phone retailer Best Buy uses analytics platform to produce business reports

Getting the measure of collecting data for a future of personalised service

With, it could be a matter of changing the sort ordering of the intimate garments sold on the site.

Ross contends that online retailing is about what he calls “inputs” that can be controlled. Amazon is, he says, “really a fulfilment business”. As an example of an input he cites page-weighted price competitiveness.

“You want to be price competitive on the top viewed products. You can action that – that is to say, you can reduce prices. Similarly, are you shipping to customers when you said you would? Again you can control that, and that is where you focus your efforts,” says Ross.

“E-commerce is more Toyota than Tesco. It's more about the industrialisation of decisions based on data. It's not about having an army of data analysts. It's about automating. Physical retail is very observable. Online you've got lots of interconnected processes that look much more like a production line”.

He draws a further parallel with the Industrial Revolution, which de-skilled craftsmen: “This stage is all about replacing knowledge workers with algorithms”.

His company, eCommera, supplies software and services for e-commerce, and he publishes a journal, Decision Intelligence. Its DynamicAction product displays, in a dashboard, some 400 specific actions.

Customers include Asda Direct, House of Fraser and Jermyn Street tailor T.M. Lewin.

“Because there are so many micro activities that happen in the e-commerce enterprise, you need hundreds of reports to make sense of them. That could be unmanageable. The reason you need a different approach is you need some way of making sense of those hundreds of reports. You need a methodology to rank the reports/actions based on predicted impact on profits,” says Ross.

“There is also the matter of drilling down from profits to the inputs to enable the leadership to understand how it all fits together”.

He argues that the retail sector is ripe for consolidation: “Look at - it's bought many businesses the last few years. You will see conglomerates emerging who are the ones who have really cracked the code. They will run retail businesses more profitably and efficiently, with massive economies of scale. Most industries have consolidated. Retail has not been like that. There is a long tail of ‘Mom-and-Pop’ shops”.

And, fundamentally, he says, “These days, retail is hard”.

Read more on Business intelligence and analytics

Data Center
Data Management