Charlie Pool
21.05.2018

data science, engineering, proptech

weightedcentrewarehouse

The map above takes our last three months historical data for requests for warehouse space that have come through the platform and marks them as red dots.

These are validated requests as opposed to a user just playing around with the platform or just fishing for a price. What is interesting is you will see the red dots in many of the places you’d expect to see them. There’s a clear clustering in London, and within London you can see a distinctive West, Central and East London cluster. Then largely the rest follow infrastructure hubs - motorways, airports and ports. There’s nothing much new here, but still it is interesting to see the data confirm what you might expect.

Where it does become interesting is when you start to segment the demand. We can do this by sector, size, services required, size of company, margin - a whole range of variables, but here we have chosen to focus on value.

When a customer puts a request for a warehouse through the Stowga platform as well as saying where they want the warehouse to be they also have to give other pieces of information such as how long they want the warehouse for, how many pallet spaces they need, and how much throughput or turnover of those pallets they expect. When they do this the platform calculates the value of their contract for the supplier.

e.g. if someone says they want to store:

  • 1,000 pallets,

  • for 10 weeks,

  • with no turnover of pallets

and a warehouse quotes:

  • £1 per pallet per week storage

  • £1 per pallet in handling, and

  • £1 per pallet out handling

The total contract will be worth £12,000

1,000 x £1 x 10 weeks = £10,000 storage

1,000 x £1 = £1,000 handling fees for moving the pallets in

1,000 x £1 = £1,000 handling fees for moving the pallets out

The round circle in the middle is the weighted centroid. The weighted centroid is the arithmetic mean centre of all of the demand but weighted by the estimated value of the demand we have calculated and recorded. Think of a triangle - the centroid is the arithmetic mean of the three points of the triangle. In this case the shape is not a triangle but the the entire country, so there are millions of points to calculate. Fortunately an algorithm does that for us and finds the centre, but then on top of that we have weighted each red dot by the value of the contract. As a result, the ‘centre’ of the country gets pulled - pulled towards where the value is.

The second image here zooms in on the location that is the weighted centroid. It falls just outside Lutterworth.

weightedcentrewarehousedetail

This might seem surprising at first glance - with all that demand in the South East focused around London you might expect the weighted centroid to be dragged towards London. It is not. Instead it is right in the heart of the golden triangle, deep in warehouse country. What that tells you is the weight of the many smaller contracts in the South East is almost perfect counter-balanced by the weight of a few big contracts in the North and Midlands.

Whilst Stowga has had demand at Magna Park which lies a few miles to the West of the centroid, that demand is not significant enough to be an anomaly that would skew the data. Lutterworth is genuinely the centre of gravity for the value of our deals. There is more demand in London but the deal size is small so does not pull the centre. The big deals are up north.

To people in the industry this probably doesn’t seem all that surprising. Everyone knows this right? Well yes, we do - but it’s interesting to see the data confirm this nonetheless. What will be interesting is to track this over time and see how the weighted centroid and the distribution of demand moves with changing seasons and changing consumer habits.

We are currently looking for a business analyst to join the team. If you enjoy these data insights please Like and Share the article - it might justify a bigger budget for our analyst!!