Interview: How algorithms, automation and ‘the milkman model’ are powering Picnic

Titans of the trade, including Albert Heijn, Jumbo, Aldi and Lidl, have had their foundations rocked by the emergence of an online food start-up using exactly that approach to fulfilment.

An efficient use of milkman-like fixed delivery routes, coupled with immense tech and digital expertise, has proved the basis for the rapid rise of Picnic, the pureplay supermarket that is gobbling up market share in its homeland.

The etailer, which only became operational in 2015, bills itself as “tech’s answer to groceries”. And it is easy to see why.

Orders can only be placed through Picnic’s smartphone app, it uses technology to automate operations in its warehouses and proprietary algorithms regularly map out and adjust the most efficient delivery routes for its drivers as new customers come on board.

Shoppers can place orders as late as 10pm for next-day delivery, which will arrive at their doorstep within a designated 20-minute window.

That unique business model has been constructed around an equally unique proposition – Picnic promises its customers the lowest prices in the Dutch grocery market and free home delivery. The only caveat? Baskets must be worth at least €25 (£22.70).

It is a model that has wooed shoppers in their droves. Picnic now boasts 250,000 customers across 80 cities in the Netherlands, expanding from Amersfoort into bigger markets including Utrecht, The Hague and Amsterdam. It also operates in 10 cities in Germany, where it debuted in Dusseldorf in March 2018 following a €100m investment the previous year.

Although heavy investment and swift expansion have kept it in the red – losses widened to €45.5m in 2017 – Picnic’s emergence has sent such shockwaves through the Dutch grocery industry that analysts in the Netherlands coined the phrase ‘the Picnic effect’.

Delivering something new

“We are becoming market leader [for online grocery sales] six months after launching in a city,” Picnic’s chief technology officer Daniel Gebler – one of the keynote speakers at next month’s Tech. festival – tells Retail Week.

“We look at how much online food consumption is increasing within the markets where we are active. This is something where we are seeing major jumps. In the Netherlands, they call it ‘the Picnic effect’ – in cities we go into, the online share of food retail is significantly increasing.”

In the Netherlands, online food penetration stands at 4%, compared to around 7% in the UK, leaving plenty of room for growth. Picnic already accounts for 23% of that online food market, compared to the 5% slice of the pie it held just three years ago in 2016. Its market share makes it the third-biggest food etailer in the Netherlands, behind only Albert Heijn and Jumbo.

“Every day we release a new version of the driver app, of the warehouse system, of the route-planning system”

Daniel Gebler, Picnic

Reaching that level, and meeting the high expectations it sets itself around both price and fulfilment, has been far from easy – particularly when Picnic’s strategy is to keep its technological innovation in-house, rather than partnering with third parties in the way Marks & Spencer and Kroger have with Ocado, or Carrefour has with Google.

“Our strategy is a challenge for tech and innovation,” Gebler admits. “How do you power such an engine? What is a delivery model that you can bring to customers that means you don’t have to ask for delivery fees?

“From the technology side, you quickly learn that if you want to be a strong logistics provider, you cannot use commoditised market solutions in logistics, in order picking, in warehouse management, in fulfilment, in purchase forecasting. You need to build this technology completely in-house.

“At the beginning, we used a standard logistics planning solution to plan deliveries to our customers, but none of those solutions are optimised for efficient inner-city logistics. These are all for long-distance logistics if you want to deliver something from Dover to Calais, for example.

“We had to come up with a completely new route-planning algorithm, that we developed ourselves, to make route planning as efficient as possible.”

Pushing boundaries

Picnic’s push on proprietary technology extends far beyond that. Its growing data science team has developed a machine learning tool to help it forecast the groceries customers will be buying “tomorrow, the day after tomorrow and next week”, Gebler says.

He explains that the technology has allowed Picnic to work more closely with its suppliers and make “much more robust decisions” around order volumes, allowing it to have “close to zero waste”. Such efficiency and stock control can then be fed back into the consumer proposition, in the form of lower prices and free deliveries.

Picnic has also built systems that make parts of its warehouse operations automated. Tote movements within the retailer’s depots are driven by technology, though staff members still manually pick and pack the products into Picnic’s fleet of electronic delivery vans.

That’s something else Gebler wants to change. He has tasked his 100-strong tech team to come up with solutions that will take the business “in the direction of Ocado”, by using robotics to allow for “automated inbound, automated outbound, automated picking, automated replenishment and automated dispatching” of groceries.

He intends to push technological boundaries in other areas, too. Gebler has put “a few engineers” through training that will help Picnic create autonomous vehicles in the future. The etailer is already “looking into options” on this front and has plans to test self-driving prototypes in the coming years.

He dubs this “a really exciting space” for the business, but admits the need for “a very, very high level of autonomous driving sophistication” – in order for vehicles to avoid hazards such as pedestrians and pets crossing roads – means self-driving vans are not a short-term Picnic priority.

The biggest challenge

How, then, does Gebler juggle the demands of such necessary future-gazing with servicing the day-to-day technological demands of the business? He highlights that trade-off as the biggest single challenge he faces in his role.

“There is much more demand and vision within the business than any form of tech team can ever realise. We have to have a careful process for choosing ideas,” Gebler says.

“Because it is an operational business, you have to think about how to balance between building new systems, new technology, new propositions, while also supporting the ongoing business.”

“I want everybody thinking about not only how to solve the problem of the day, but  what Picnic should look like in a months’ time, a year’s time, 10 years’ time”

Daniel Gebler, Picnic

He says Picnic aims to have its tech team spend at least half of its time building “completely new systems and propositions for customers.”

“We are iterating on usually a daily or weekly basis, making small improvements. Every day we release a new version of the driver app, of the warehouse system, of the route-planning system,” he says.

“But what everybody should be aware of is that there is so much more ahead of us than behind us. I want everybody thinking about not only how to solve the problem of the day, but thinking about what Picnic should look like in a month’s time, in a year’s time, in 10 years’ time. If you have that kind of mindset, you come up with solutions that tackle short-term problems and that can make drastic improvements in the long term.”

If Picnic’s technological prowess can achieve that, it will give grocers across Europe even more food for thought.