90 Days to Usable Data
Here’s what’s holding your data-driven marketing back — and here’s what to do about it.
The goal of Pat’s Pantry is simple — give people time to enjoy a healthy meal at home. To this fictitious on-demand meal kit service, it’s not about the actual food they deliver — although that’s important. Instead, it’s about the experience of enjoying a meal with the people you love — easily selecting meals that meet your dietary requirements, having food delivered when it’s convenient for you, making food prep as simple as possible while still achieving a warm, home-cooked meal. The experience of being a Pat’s Pantry customer should be as enjoyable as the meal is.
To deliver these memorable, meaningful experiences that exceed what the competition offers, Pat’s doesn’t just need fresh produce and meats — Pat’s needs data. And that’s where the challenge begins. While Pat’s has data on everything from who the customer is to where they live to what foods they eat, don’t eat, and can’t eat, trying to understand it in a cohesive, usable way has taken the brand down. It’s time-consuming, resource-heavy, and the pieces never quite come together.
Data counts — and everyone knows it
Pat’s is not alone with its data woes. Big data has never been bigger — literally. Within that data lies your brand’s ability to deliver unparalleled customer experiences at scale — experiences customers want, and experiences that will be your brand’s competitive differentiator now and in the future.
“The ability to create experiences that are relevant, personal, and timely, via moments that assist the customer through their journey, is a priority across the entire enterprise,” says John Bates, director of product management for Adobe. “Any person who owns any of those moments needs the right information and context to make them magical.”
And therein lies the key — you need the right information for the moment you own.
Too often, marketers are enticed with the promise of what can happen when all data is unified. But, in reality, to be most effective, you need to start with a business problem you want to solve and then determine the data you need to solve that problem. This gives you a manageable project and starting point on your journey to becoming a fully data-driven experience maker.
“Identify one or two needs where you have the greatest opportunity to leverage data to drive action — and then measure the success,” John says. “That’s going to inspire alignment in people throughout your organization. It helps to secure buy-in from executives so they can push these initiatives from the top down and it gives justification to the long-term effort.”
Here, then, is a process for applying data to your business problems so you can deliver the experiences that will add value for your customers — and your business.
Step 1: Establish expectations and business purpose (20 days)
For Pat’s, one thing they want to do to achieve their goal of business growth is acquire more customers. To solve for this problem they want to know who their best customers are and how to find more like them.
In the first 20 days of their journey, the new customer team at Pat’s has identified a business need they want to solve with data, established expectations, and gained executive buy-in.
Step 2: Gather the data (8 days)
Next, the team at Pat’s needs to gather the necessary data. It’s essential to understand the data they need and where it’s coming from. Pat’s has data on a variety of platforms:
- Online behavioral data including customer sign-ups and opt-ins from an ordering app or web interface
- In-house tracking of order and delivery history, including location and food quality
- Internal management dashboard with intel such as status reporting and alerting
- Product and supply data such as food and ingredient sourcing
- Third-party customer satisfaction data from an outside market research firm
- Content Management System (CMS) data culled from content creation spanning mobile app and website
- Email campaign data
- Ad network display campaigns
- Social promotions
- Customer relationship management (CRM) data for corporate clients and large accounts
- Data management platform (DMP) with first-, second-, and third-party audience data sets
It’s a lot of data coming from a lot of disparate sources — but that’s not uncommon. Even for a small B2C business, there are a tremendous number of data points, insights, and observations bubbling up 24/7. And, again, capturing them is just half the battle.
Pat’s determines a few different data sources that will be most helpful in finding out who the best customers are and where to get more of them. The acquisition team decides to look at in-house tracking of order and delivery history, third-party customer satisfaction data, online behavioral data, DMP audience data, and data from email, ad, and social promotions.
Step 3: Ensure privacy through targeted data governance (15 days)
Pat’s is realizing that many of its data sources have restrictions. For example, some of the second-party data in its DMP was provided under the condition that it will not be used for off-site targeting.
In reviewing your data collection and usage workflows, ensure you’re in sync with current internal and external regulations and data privacy laws. During this assessment it’s also important to understand new regulations and policies coming to the surface, so your organization can protect itself from future liability.
This future-proofing is built into Adobe Experience Platform, enabling customers to roll with changing privacy standards and environments, while still successfully collecting, managing, and leveraging data.
With the recent move toward more customer-centric privacy rights, it’s essential to take time to gut-check workflows and processes. “Once you’ve closed any privacy gaps that exist,” John says, “you’re just about there.”
Step 4: Map what’s missing and prioritize data needs (20 days)
Now that Pat’s knows the data sources it will tap for its customer acquisition plan, they need to identify any content-tracking variables that may be collecting inconsistent information. This can be accomplished with a tag management and a tag audit.
During this process they’ll likely identify areas where data collection falls short. “If data doesn’t share a common data field across data sets — or there isn’t a common key — then take the needed steps to get that data created and collected appropriately,” John says. “It’s never a wasted effort.”
Pat’s determined the third-party customer satisfaction data it has doesn’t integrate easily with it’s other data sources, and it already knows it has limits on the partner data it has access to. With that knowledge, Pat’s will de-prioritize this data and also work with the customer satisfaction vendor to improve future collection.
It’s also critical to ensure common keys are created to link AR and VR data, mobile data, voice-based device data, and other IoT data sources to other intel and insights. Understanding these points will impact how future data is collected and, with it, your workflows and processes going forward.
“Don’t get bogged down, though” John warns. “Prioritize your data needs based on how critical they are for the business purposes you established in step one, and record the value each dataset provides, relative to the effort required to connect it.”
Step 5: Stitch data to create real-time customer profiles (12 days)
Get to this stage and you still likely have more data than you can handle on your own. Pat’s is going solve this obstacle by tapping into a common data model to stitch together the transactional, behavioral, financial, and operational data it has into a single, actionable, and holistic customer view.
“The collection, the cleansing, and the processing of data can all be aided with AI,” John says. “Through this process we get rid of about 80 percent of the work a data scientist would have applied before he or she even got to start on the fun stuff.”
As Pat’s stitched its data to create real-time customer profiles, they improved data governance and streamlined and simplified experience delivery. And most importantly, they were able to determine that their best customers usually order three meals per week for two people, live in urban areas, order from a mobile device, and use discounts from social ads.
Pat’s now has verified reasons to invest more in its mobile site, spend more on social promotions, and provide more meal options with two servings.
Step 6: Build a roadmap to use data for more business questions (15 days)
Having applied data successfully to one business problem, Pat’s is ready to flesh out a comprehensive roadmap of the processes and technologies needed to set the wheels in motion for more data-driven experiences.
“Have a path in place whether that’s to update your technology or your internal processes,” John says. “Then go into action and learn from it.”
To do this in your organization, refer back to step one, and apply data to your business needs one step at a time. From there, you’ll be able to keep moving forward, tackling challenges and goals as a data-driven organization.
For Pat’s Pantry, they’re moving forward to determine the benefit of offering regional meals by looking at whether Pacific Northwest customers order different meal kits than customers in Florida. They also want to understand the commonalities between their most valuable customers — what they like, what they don’t like, order frequencies, etc. — so they can customize offers and packages that exceed expectations.
Identifying — and accelerating — use cases
Now it’s time for Pat’s Pantry to roll out the full-scale data integration. In Pat’s case, they’ll turn on server-side forwarding between their analytics and DMP for real-time audience enrichment. They’ll also leverage the Adobe Analytics and Adobe Target integration for integrated A/B testing and targeting campaigns.
That said, keep in mind data is only as good as the value it delivers — and the value your stakeholders see in it. “To show the value of data without question, you need to nail the specific key results you will use to measure success against a business objective,” John advises. “Demonstrate the degree to which you can pull together data and then execute against one use case, rather than trying to boil the whole ocean for stakeholders.”
“After that, optimize,” John says. “Uncover key areas of your customer experience to improve. Design A/B tests that minimize bias and maximize test efficacy measurement.”
By looking at the data and reporting, Pat’s data analyst noticed Pacific Northwest customers prefer fish-based meals. Armed with that information, they can create a test campaign to maximize subscriptions for prospects in the Pacific Northwest — testing hero banner imagery with different fish-based meals. They could also easily test food categories like Indian, Coastal Carolina, and comfort. Tapping into customer preferences helps deliver a better meal experience for its customers, while also growing revenue for Pat’s Pantry
“And, beyond all of this, don’t forget to share insights and evangelize the success across your business,” John says. “Keep the momentum going and keep everyone on board with your data fitness. Do that and you’ll be able to easily expand to other stakeholders.”
Read more about enterprise organizations putting their data to work in The Data Rush: How to Strike it Rich series.