Customer data platforms (CDPs) are popping up left and right. (Check out our simple definition, if this is your first time hearing of the term).
We interviewed the Director of Analytics at Just Salad to go straight to the source and find out what it’s like to evaluate and implement a customer data platform for an enterprise brand.
If you prefer to skim written content than listen to and watch a video, we’re here for you. Here are the top questions we covered — and Or’el’s answers. summarized:
A CDP is a system that ingests data from multiple sources. The pain point that a CDP mainly solves for is that customer data lives in many different places, and each source does not connect to another, so by definition, you need at least two sources.
What a CDP offers to a business is basically a unified customer view. Within a customer data platform, an anchor for any data structure is unique identifier associated with the customer (i.e. credit card number, email address, phone number). This gives a business the ability to follow each customer through different channels across lifetime and share insights associated with that
Here's a visual "Before" example of what an enterprise restaurant brand’s data flow might look like before a CDP:
And here’s an example of an “After” picture, WITH a CDP:
(Want more information on what a CDP is in the first place? Check out our blog post!)
We talked about the problem of data silos, but let’s dig into the impact. Often a brand is forced to make business decisions off of one source of data versus a combination of all sources. So, for example, if you have 20% of your customers enrolled into a loyalty program — which is a very successful number, often times brands have closer to 5 - 10% of customers enrolled — even then, that’s a small subset of your customers.
If you’re using data from 20% of customers to inform decisions for 100% of guests, you’re leaving opportunity on the table.
We didn’t have access to see or easily measure Customer Lifetime Value (CLV) in one place. We needed to know CLV so we could have a baseline and measure changes over time. It’s only with the ability to see CLV that you can pull levers to increase it (i.e. frequency, recency, average order value). It’s essential to know where you’re at to start and to have the ability to automatically monitor changes over time so you can actually see if you’re improving CLV and track the rate of improvement over time at each location.
Yes, this cannot be a business decision that’s made in a vacuum of one team. When we first started having conversations about a CDP, our marketing was like a black box. We couldn’t easily understand efficiencies, and this impacted teams like Finance, Operations, Engineering, and Data Analytics. Many of the stakeholders who weren’t as close to the problem day-to-day initially saw the ask for a CDP as an added line item. Just another vendor cost.
But then with internal education, internal stakeholders began to realize the importance of a CDP as a connector. To this day, we now refer to our CDP as “the connector.” To convince the c-suite to move forward, you need to get people to understand both the opportunity cost of not having a CDP and every opportunity for actionability of the data. Lay out every use case for what you could accomplish in the brave new world of having a CDP, and paint the picture for other teams. It can help to run through use cases of what you can do with a clear sense of cohorts, recency, frequency, and customer lifetime value. You just need to get over the hurdle of lack of alignment.
Sure, let’s cover two examples.
One is more intuitive. Let’s say you have a cohort of customers that demonstrate a really valuable behavior to your brand. How can your brand get more of that behavior from other customers? To answer this question, you’re either going to be deep in spreadsheets doing a one-time look into your data, or, if you have a CDP with BI tooling built in, you can log in to your platform and access this in seconds. Let’s say for Just Salad that you realize that the cohort of customers who order a higher-priced salad are more likely to order other add-ons as part of their meal (drink, etc.). To drive up AOV with other customers by nudging them into this cohort, you could run a promotion on that higher-priced salad to get more customers to purchase it.
Now let’s quickly talk about a second example. This year, Just Salad sponsored the Brooklyn Marathon. We were able to approach this business decision with confidence because of our CDP. We were operating not off of gut but off of data. Remember, data always wins over opinion. At the leadership level, we looked at the number of race attendees, could predict how many impressions we would get with a promo, and knew the conversion rate from impression to initial purchase as well as the CLV of our customers, thanks to our CDP. From this we were able to predict exactly how many times new customers needed to order to justify our investment. We weren’t driving blind, and we had the proof we needed to move forward with sponsorship.
Sure, let’s go back to the Brooklyn Marathon example. Marketing could look at the cohort of guests who made first purchases based on impressions at the marathon. They were able to identify impressions at the marathon as a brand touchpoint that led these customers to place their first order. This is huge. They could also track through to see if customers dropped off after that first purchase, within two weeks. For those guests who did, Market was able to trigger a reactivation campaign. Looking at this in dollars and cents, the CDP allowed us to maximize our investment into the marathon promotion and both create and recapture customers. One finding that was cool through a menu item analysis via our CDP was that we found that marathon runners have healthier diets, and so now we’re able to run personalized campaigns to these guests around healthier menu selection.
CDPs are interesting because the category creators like Segment aren’t specific to any one industry. They are horizontal, so they don’t have things that Bikky has, like POS integrations built out of the box. Because of this, we had to get creative with point of sale data. We needed our offline data to be pulled into our CDP — both retroactively and into the future. So we ended up using order ID, purchaser information, and tokenized credit card data to match each guest to their online transactions and marketing engagement. It’s important that you build a legitimate contact record (with marketable information like email address and with order data that includes how frequently the guest visits, how recent they last ordered, et cetera.
If you're in the early stages of considering a CDP, there are three high-level, immediate benefits to keep in mind for your business:
>> Hungry for more CDP content? See our full introduction into what a restaurant CDP is.
Big thank you to Or’el Anbar for guest-speaking with our CEO, Abhinav!