COVID: Cloud or .... Silver lining?

In conversation with a customer, I ask - subtle and direct as always - whether they see the benefits of Corona as well as we do. Silence at once.

I say: you probably think “what is that fool talking about? To ask this from a retailer in this day and age”?! Now it happened to be a hardware store that - not unimportant – experienced the best year ever last year! But at that moment I was not aware of that.

“Hmmm” the customer mumbled, “a remarkable question indeed”. However, definitely not a surprising question for researchers, consultants and developers. Since for them “Good and Bad, Crazy, Weird or Inappropriate” are just opinions and interpretations. And therefore not (so) relevant. 'Problems' are data to work with. And that makes it all the more interesting!

Retail calculation

Earlier, singlebeam infrared systems where widely used. Using lightbeams counting when interruption occurred by walking by. End of day counts were divided by 2 for daily totals. With these numbers the conversion was then calculated to make comparisons on various points.

But real mathematicians question: why divide by 2 if that is done everywhere? Makes no sense! Check for yourself: 4%, 6%, 14% (conversions from 3 locations) or 8%, 12%, 28% both give the same ratio: 2 vs 3 vs  7. Dividing by 2 feels better as better approach of the actual number of visitors. But...: how scientific is 'feels better'? And even more interesting: do you really need the actual number of visitors for (conversion) calculations?

Come and shop alone!
Logical and simple corona rule, right? But one with major consequences! A commonly used KPI changes drastically! And with it all kinds of derivative relationships. If you don't take this into account, you will make wrong decisions. Date rules! But only if properly interpreted...

Hardly any supermarket kept track of it’s number of visitors. Everyone who enters the supermarket also buys, right? So conversion is still around 100%? Nope, wrong. The conversion is between 45% and 87%. How come? Easy. School children often 'shop' during their breaks. Not all group members always buy something then. And many people hate shopping. Me too, so I shop with my wife. That makes two people with one wallet. Thus dropping conversion to 50%! Only: we form one buying unit together only.

Supermarket conversion is very high on buying units. Not so on the number of visitors. And specificly those are counted. Groupsize of the buying unit also changes with time and day. Influencing directly –but wrongly– your score. This is the biggest effect of the “Come alone” rule!
So: actually do not measure numbers of visitors, but numbers of buying units! Often possible with the same sensors. As long as they are set up and placed correctly.

flow
Another interesting formula: Throughput (=#visitors\ average dwell time). Revenue is generated by numbers of people through your store. Taking time for buying moments. If Social Distance rules restrict this very much, this formula becomes very valuable! When max. 'number of customers in' is reached, queues must be avoided! Waiting too long = running away.

                                 #visitors
Throughput= ----------------------------
                        average dwell  time

remains high with many visitors. But that is unfortunately a limitation now. But can be maintained by limiting 'average dwell time'. So when the store is filled fully, help visitors to find what they are looking for and do not waste time roaming around in the store. To keep waiting times at the entrance short. To avoid people run away.

Testing your feeling permanently with numbers: there is no better starting moment!