in collaboration with dunnhumby

The Challenge

dunnhumby manage and analyse large transaction datasets and we are constantly looking to improve the speed of extracting customer insight.

Many metrics we calculate are non-aggregable. For example, customer penetration is the number of customers who bought a specific product divided by the total number of customers who shopped.

  • Example: customer penetration for Easter eggs = number of distinct customers buying any Easter eggs divided by number of distinct customers buying any product.
  • As this metric involves calculating distinct customers from a large amount of raw data, it cannot be pre-summarised; consequently it can take a long time / be expensive to calculate.

The challenge is to find ways to calculate non-aggregable metrics – such as customer penetration or similar – on large data volumes in less than one second, using specialised processing that target, for example, hardware architectures based on FPGAs.

We are pleased to co-fund a PhD scholarship with Imperial College in this research area. The successful applicant will collaborate closely with dunnhumby’s computer science team, who are based in our Hammersmith headquarters in London.


dunnhumby is a world-leading customer science company. We analyse data and apply insights from nearly one billion shoppers across the globe to create personalised customer experiences in digital, mobile, and retail environments.

dunnhumby employs over 2,000 experts in offices throughout Europe, Asia, Africa, and the Americas and works with a prestigious group of companies including Tesco, Monoprix, Raley’s, Macy’s, Coca-Cola, Procter & Gamble, and PepsiCo.

How to apply

Those who wish to be considered for the PhD scholarship should send Professor Wayne Luk ( their CV and a one-page summary of their experience and their research interests, explaining their relevance to the project described above, preferably by 1 May 2017. Informal enquiries about the proposed project can also be sent to Professor Luk.