Case Study
.png)
co-written with Cezar Rachieru, VP Operations, Duffl
Duffl (www.duffl.com) delivers snacks to college students in 10 minutes. They buy wholesale high demand/high margin products and deliver them directly to customers using electric scooters and a student-driven workforce. They’ve launched at: UCLA, USC, UCSB, UC Berkeley, UofA, ASU, TSU and UT Austin.
Cezar Rachieru, who leads Duffl’s Operations team, had a unique challenge on his hands. He operates 7 campuses, selling 1000s of SKUs for a very price sensitive customer base in a deeply competitive market.
He has a very seasonal operational business. For example, willingness to pay for Red Bull in exam weeks is different than during spring break, or ice cream is more popular in the August than it is in January. Cezar realized that these seasonal patterns should be incorporated into his business decisions, but how?
Cezar knew that his well funded competitors were deploying data science teams to build pricing algorithms that his startup team couldn’t build in house. That’s when he decided to work with Luca –
We were in a catch 22 situation – if you over optimize for your profitability, you shrink your business. But at the same time, you need a business where the unit economics work. We needed a core pricing solution that could execute on our pricing strategy and account for environmental factors and considerations that we just candidly couldn’t manage ourselves. Enter Luca. -- Cezar Rachieru, VP Ops Duffl
.png)
The Duffl team cared about sustainable growth above all else. They configured Luca’s price optimization engine to meet a certain campus-level margin goal, but gave it flexibility to grow revenue within that margin constraint.
They also really cared about protecting their users from large swings in prices, too frequently. So they put in sensible constraints around how frequently prices could be changed or how large each price change was allowed to be.
As long as those constraints were respected, the Duffl team permitted the Luca system to find arbitrage and optimization opportunities.
The Duffl Team decided they needed to get the humans out of the loop, and make the system as close to automated as possible. By integrating wiht Luca's API, the Duffl team was able ot transmit real-time sales data and push price changes through Luca's dashboard. This way, the Duffl operators could get out of the way of each price change, but still be involved in an approval process if required.
Within 4 weeks of the Luca integration, Duffl saw 2% in revenue growth and 11% of profitability growth, as well as started to build a deeper understanding of their business.
Before Luca, our pricing processes were time consuming and extremely manual, without any science to power it. We were simply starting with an idea and observing what happened without the tools to validate or invalidate the output. When we deployed Luca’s tech, we were able to understand what was happening in our business at a SKU level. -- Cezar Rachieru, VP Ops Duffl
We asked Cezar how he was able to get comfortable with outsourcing pricing to an AI system. This was a process his team was used to running manually, so why adopt a solution like Luca to replace it?
This is what he had to say –
As a startup, It’s important to understand your own resource constraints. When something is as core to your business as pricing is to ours, you’ve got to think hard about what you’re leaving on the table through outdated manual processes. So if there is a better choice, meaning there is a system that can protect you, where you can put some rules in place and then let the system optimize on your behalf and reach much better conclusions than you can, then it's a no-brainer. It's really what technology is supposed to do. -- Cezar Rachieru, VP Ops Duffl
We couldn’t agree more, Cezar.
