Four times better e-commerce logistics through use of big data

By making smart use of data, e-commerce companies can
significantly improve their logistics operations. That is the view of Jack Pool
of consultancy firm Districon. “There’s a lot of data available, but the question is how
we can make good use of all that data.” At a conference organised by the University of
Groningen (RUG), Pool gave four interesting examples.

Retailers and e-commerce companies are struggling with data,
says Jack Pool. “We received an enquiry from a company with a central
distribution centre in Europe, which wants to deliver to its customers much more quickly
. This can be achieved by setting up warehouses in major cities, but which
items do you store there? That’s also the dilemma facing Amazon, which has
has set up a warehouse in an office block. That warehouse is small, but
Amazon has a massive product range. So what proportion of that do you stock in
that warehouse in Manhattan? To answer that, you need to use data analysis.”

, Pool, gives fourexamples of specific cases in which
data analysis has proved to be of great value:

Case 1: Dropshipping.
Districon investigated on behalf of a retailer which proportion
of the product range could be managed without holding stock in the retailer’s own distribution centre
but instead to be held by the supplier. These turned out to be mainly
items with a low turnover rate and a high value. By
implementation, the retailer has achieved a cost saving of 28 per cent
.

Case 2: 3D slotting
Most companies only look a few times a year
at the location of fast-moving and slow-moving items in the warehouse. What they
forget to take into account the height of the picking locations. An item
located on the highest or lowest level of a shelving unit takes longer to pick because
takes longer to collect because of the need to reach up or bend down. By taking this into account
, cost savings of 5 to 10 per cent are possible.

Case 3: Delivery from shops
For a retailer with an e-commerce distribution centre in Amsterdam,
we investigated whether it is feasible to deliver to consumers from shops.
particular for more remote regions, this could be an attractive option. Analysis shows
that it would be worthwhile to adapt fifteen to twenty of the one hundred shops
to handle the dispatch of online orders. This has now
been implemented.

Case 4: Giving consumers a choice
Research shows that ten to twenty per cent of
consumers would opt for a delivery date later than tomorrow, if they were
option is made available to them. Applying this concept to various
scenarios shows that savings of 9 to 31 per cent on labour costs are possible
as processes in the distribution centre become more predictable and peaks are
smoothed out.

Leaders go further
To illustrate the importance of data analysis,
Pool cites Gartner’s data model. This model distinguishes four stages in
data analysis:

1. Descriptive analytics: what
has happened?

2. Diagnostic analytics:
why did that happen?

3. Predictive analytics: what
is going to happen?

4. Prescriptive analytics: how
can we ensure that something happens?

“Most companies haven’t yet progressed beyond the first
two stages; sometimes they do a bit of forecasting,” says Pool. “The frontrunners
, on the other hand,
do go further and invest in
centres of excellence in the field of analytics.”


Marcel te Lindert – Logistics and supply chain journalist