The full package: How Evri is using advanced data analytics for happier couriers and customers
“Our parcel scans alone are well over 100Gb a day”
Evri’s first Chief Data Officer explains how the company is using a new data analytics platform to deliver more efficient service, grow the business and meet ESG targets.
Harvinder Atwal is Chief Data Officer at Evri and is the first person to hold that post within the business.
“There are two dimensions to scaling at Evri,” says Atwal. “One is the long-term growth of the business. The business has been growing at an incredible rate for last 5 – 10 years so any data platform must be scalable long term.”
Then there’s seasonal scaling.
“Peak in the industry runs from Black Friday to Christmas, and we see significant jumps in in volume during that period.”
It’s worth taking a moment to appreciate the scale of this peak. Evri announced earlier this week that the week following Black Friday 2024 was its biggest week ever, with 24.7 million parcels being handled. Total parcel volumes rose by 12% to 173 million over the nine weeks to 28 December 2024. On average the courier delivered 2.4 million parcels a day in the run up to Christmas and on its busiest day (December 4th) it delivered 3.8 million.
Each one of these parcels generates data at multiple journey points.
“We need to be able to handle lots of different data types,” Atwal says. “Our primary source of data is streaming data. Every time a parcel gets scanned, either physically or via a machine process, it generates an event track point. Our parcel scans alone are well over 100 gigabytes a day.
“On top of that we have operational data like our transport systems. Our vehicles have trackers on them and that all transmits data. Then there’s people resource data such as time and attendance, financial and location data.”
Evri chose Databricks as their data partner for several reasons. Atwal explains:
“Databricks had the ability to handle large amounts of near real time data that we that we have - at scale - and process it very quickly. Our use cases weren't just traditional data warehousing. We knew that the journey was going to take us towards more advanced analytics and machine learning and AI and Databricks could handle that without needing a separate ecosystem of technology around it. We were also impressed with their development roadmap.”
Courier Pay
“Contrary to what people might think, it’s not in anyone's interest to underpay couriers,” Atwal says.
Smart data analytics have provided a much clearer view of what fair pay looks like, and helped couriers understand what they will earn over a certain time and/or route.
“We need to know many miles did they travel? How long did it take them? We need some information on the routes they took. We also bring in more adjustments for local factors, cost of living in different areas. We bring all that data together on the platform to make sure that the pay rates we set are competitive.
“Previously, it would have been really hard to do that kind of modelling at scale, to understand the 3 million deliveries we make every day across those 25,000 rounds, just exactly how much time it took between stops and the distance travelled by a courier.”
Future and skills
How does Atwal see the next 12 – 18 months shaping up at Evri? The company aims to be delivering a billion parcels annually within five years.
“Our plan is to continuously increase the amount of data sources because a lot of the value we can get from our data is in being able to integrate lots of different data sources.
“Then it's about accelerating use cases. Self-service analytics are a big objective. Lots of people are reliant on reports to run the operation day-to-day. We don't want the data teams to be a bottleneck, so we want to democratise analytics.”
Avoiding that bottleneck involves a scheme to roll out data apprenticeships including everything from data science, data analytics through to data engineering.
“I'm keen on getting these apprenticeships expanding so that we have lots of data literate people throughout the organization,” Awal says.
In addition to everyday service and operation improvements, other use cases involve being more predictive in areas such as loss and claims, so that rather than reacting after a parcel has gone missing or been damaged, Evri can intervene earlier.
Atwal also mentions ESG, and how better analytics can help Evri hit its goal of carbon neutrality by 2035.
“A lot of our corporate clients want more granular data around carbon impact and when we're rolling out zero emission vehicles, we need to measure the impact. We want to measure at a more granular level to see the CO2 impact of individual parcels.
“We need to gather more data to do that so that will be part of our roadmap along with better service and operational efficiency is also part of the plan driven by better forecasts and better predictive analytics.
“There's a lot of scope in our out of home network as well - parcelshops and lockers. We’re continuously looking at how we optimize those locations as well as the prospect of international expansion and fulfilment where we manage the stock and the delivery on behalf of a client.”