Hackathon: Using Technology to Uncover Unique Data Solutions for Lou Malnati’s

Now let’s talk about improving pizza operations and what our professionals were able to uncover using data and technology.

Lou Malnati’s makes great pizza, and they love their customers. But with the pandemic, all of the previous customer engagement models went out the window, and a whole new type of order fulfillment method was created.

Prime TSR worked closely with the Lou Malnati’s management team to come up with complex questions to challenge the hackathon teams.

The questions were:

1. How can I get a customer to order pizza just one more time per year?
2. How can I improve the new curbside fulfillment method?

Here’s how we set up the hackathon:

1. Break the 45 employees from all disciplines in Prime TSR into two teams and give them access to the data.
2. Give them both questions and turn them loose.
3. Each team further split tasks—one sub-team focused on curbside pickup, and one focused on data/analytics.

The results:

How do we get a customer to come back just one more time during the year?

So, if a customer typically orders twice a year, our goal was to incentivize them to order three times a year.

The data/analytics teams combined all of the data from loyalty, sales, and online ordering.

What they uncovered is that the loyalty program helps with recency, frequency, and value. In fact, when an online customer redeems a loyalty reward, 52% return within 60 days after redemption—which is 106 days faster than the average customer.

Our teams also recommended ways to make the loyalty rewards program more effective, including increasing awareness of the program and making it easy to enroll and use via all appropriate channels. Having strong reporting and metrics will also be very important when measuring business value and effectiveness.

Improving the customer experience of curbside pickup

It’s no surprise people like getting their pizza without leaving their car. Lou Malnati’s (and many restaurants) had to repurpose staff, create new job roles, and develop new processes to support curbside pickup.

There were a lot of challenges to this new pickup method:

• Run out to the cars
• Ask them their name
• Run back / radio back to check on their order
• Run back and tell them their status
• Make customer wait
• Run the order back out to the correct car.

The curbside team from Prime TSR dreamed up a way to use ML and streamlined processes to improve the customer experience while reducing employee costs.

The process worked like this – 

• Collect make/model/color of cars during the order-taking process.
• Install cameras to take pictures of the cars pulling into the parking lot.
• Use a Machine Learning model to match the orders to the newly-arrived cars.
• Send matched customers an SMS welcome message.
• Either run the pizza out or send another message with the ETA.

This prototype model increased the speed of pickup orders and allowed Lou's to serve a whole parking lot of cars with only one employee.

Art of the possible: What if we had two months instead of two days?

Having an entire company solving a single set of problems for two days is a good amount of time to diagnose and hack together a solution. But, it’s still not enough.

If we had two months (or longer), here’s what we think we could accomplish:

Make the loyalty program even more effective

The loyalty program works. The data doesn’t lie. Now the question becomes, how do we expand on the program? Is it a UX improvement? Is it making it available across more channels? If we had more time, we would create a real AI model to help their digital properties become smarter and personalize orders to individual customers.

A proper, smarter curbside check-in prototype

We would have loved to create a real prototype of a smart curbside check-in process with a real-time visualization. Imagine pulling into the parking lot, getting automatically checked in, and being assigned a parking spot.

A dynamic restaurant location selection solution

There are many Lou Malnati’s locations, but just like any restaurant with multiple locations, some locations are busier than others during certain times of the day.

Using the data we have, we could create a back-end solution that automatically chooses which restaurant customers would order delivery from, even if it isn’t the closest location to them. This would help balance efforts as needed.