By Guy Ainsley
To pinch a line from a recent Oxford Retail Futures Conference, “data, data everywhere”. Since the start of digital business over 20 years ago, we have grown familiar with business data to the point where the message often gets lost.
Recently, a new trend has emerged. Big data is the collection of very large amounts of data, often mixing multiple sources, and the filtering of that data into useable information.
Being cynical, this is the management consultant community catching up with Tesco’s Clubcard.
However, according to research from the London College of Fashion, businesses in fashion retail that have implemented effective Big Data processes see gross margin improvements of 5 – 40% depending on product type.
How is this done and is this margin improvement achievable in automotive retailing?
In order to make big data work, three key ingredients are needed. The first is large amounts of data - this makes the results valid and rich. The second is data relevance - the right data to get the business improvement the user wants.
The third is the right filter which processes the data into useable and balanced information that businesses can act upon.
Automotive retailers have been using some Big Data for years. Think about parts stock profiling and auto-ordering for example. But the areas where automotive retailers can really win are still not getting the full benefits.
Viper green, petrol, four door demo anyone?
New car stock pipeline is initially dominated by “launch stock” and its inevitable hangover which can significantly affect pipeline mix for years.
After launch, pipeline mix is influenced by profit issues at the factory, (stuffing stock with high profit options that customers often don’t see the value of), range updates, (resulting in premature obsolescence), other market demands and then run-out.
So, in a typical six year new car lifecycle, pipeline never truly reflects raw market demand - more a case of hoping “the public wants what the public gets”.
Big data techniques could be used to gather much wider sets of data to improve the pipeline.
For example, big data could record dealer retained margins, time in stock, sales channels used - not by retailer but by model variant.
This is possible by pooling anonymous deal file data from dealer management systems (DMS) with search information and other data. Manufacturers, retailers and customers would all benefit. Bid data fashion retailers win by supplying the right stock at the right time in the right quantity.
Used car stock profiles would naturally improve if the new car pipeline improved. After that, the used car market is already served with big data tools although these are often “race-to-the-bottom” tools that encourage dealers to, for example, price lower get higher up on Auto Trader.
A potentially more profitable way to use big data is to manage stock buying to optimise, unit-by-unit, purchase price, stock turn and stock depreciation risk. Big data service businesses manage variable resource yield very effectively. Used car departments could also.
Service departments could use big data to assess the demand for extended and flexible workshop opening hours by using big data from other retail sectors and the leisure industry in a given market.
Big data retailers in other sectors predict, monitor and respond to customer demands quickly. Automotive retailers offer the same customer service options irrespective of customer demand.
What are the risks?
There are risks and issues.
For example, someone needs to decide how data is filtered and reported on. When managers specify their own “management reports” or use someone else’s, they are biased towards their own business needs, preferences and prejudices. If manufacturers and retailers are to use big data to its full potential, agreement is needed on how it is reported so a balanced view and business benefit is achieved.
Another risk relates to the motivation of those tasked with collecting raw data. A key differences between Tesco or Next and automotive retailers is that much of the former’s data is collected automatically.
To make big data work in automotive retail, better data collection in general is needed - that means better motivation and collection methods.
So, is significant margin improvement possible using big data? Based on the low start point and proven processes used in other retail sectors, it's very likely.