Moved from 72-hour
data refresh cycle
to 14 seconds
Establish unified customer profiles across
Deliver customer experiences in real time
Break down offline and en línea data aislamiento
Consolidated customer data into 970+ million unique profiles
Sped data refresh cycle from 24-72 hours to 10-14 seconds
Streamed online and offline data into
one central repository
Engineering Manager, Information Data Services, Adobe
Customers exist in the now. They are looking to solve today’s challenges and to prepare for tomorrow. But the data that backs their digital experiences is days old, if not weeks. Who your customers were a week ago is not who they are today. In fact, as the seconds tick by, they might not even be your customer anymore.
This is what we set out to solve with Adobe Experience Platform.
When we first implemented Adobe Experience Platform in 2019, we knew who our customers were — with a 24- to 72-hour lag. Even though we prioritized customer experiences, we were still missing the up-to-date data we needed to be truly relevant. Our customer profiles were robust. Our targeting was accurate. Our personalization was always improving. But without the element of time working in our favor, our hard work didn’t always pay off. And there wasn’t a single product in the market that could do everything we needed to get us up to speed.
So we built our own. As Customer Zero for Adobe Experience Platform, we reconceived what “amazing customer experiences” actually meant. The right experience at the wrong time was still the wrong experience. With every new test and iteration, we proved to ourselves that time had been the missing ingredient our customer experiences needed to reach their full potential. Adobe Experience Platform has allowed us to reconcile the pace of data with the pace of our customers — and provide even better customer experiences in doing so.
“Adobe Experience Platform aims to be a single data platform that can orchestrate the consumption and production of data in one flow,” said Brian Block, senior manager of information data services at Adobe.
If delivering real-time customer experiences were easy, it would have already become a common practice. But the reality of real time is highly complex. From collecting to vetting to unifying data and beyond, there are dozens of practical impediments that make potential solutions seem more like science fiction than data science. Add to that the advances that have already been made in respect to fast data — most platforms update every one to three days — and it could seem like asking for more was flying too close to the sun.
But that hasn’t stopped organizations from trying. Real-time experiences have been nearly within reach for several years. The problem is that getting there has been clunky, involving a host of different products, cobbled together into solutions that never quite made the grade. As Block said, “None of them promise a complete, orchestrated multichannel offering with a common data solution, all together in one. Data coming from different sources is hard to stitch together.” Makeshift options offered a stopgap answer, but they lacked the effectiveness and operational efficiency of a single solution.
As the industry’s most vocal advocate for customer experience, Adobe was not willing to stand by while we — and our customers — struggled with imperfect answers to the real-time question. The solution that would become Adobe Experience Platform started from a place of curiosity. Could we, Adobe, create a single offering that would change the nature of our customers’ experiences? And would this solution become something that other organizations could benefit from? Per Block, “Our vision was to strike the right balance between what our business needed to run and operate, and where we needed to develop quickly to add business value.” As we tested Adobe Experience Platform for ourselves, the challenges of real-time became increasingly clear — as well as the path forward.
Engineering Manager, Information Data Services, Adobe
“Our big problem was that we had fragmented aislamiento of profile data sets. And that we couldn’t leverage the real-time purchase events coming from our own purchase applications,” said Block.
One of the biggest hurdles to real-time experiences was designating a single repository for all data, en línea and off. En Línea data, like sitio web behavior and browsing habits, already posed its own set of challenges in terms of unifying all data sources. Adobe Audience Manager already had sitio web streaming capacity for customer profiles, but there was no coordination between it and other Adobe data stores. And on the other side, offline behavior — such as transactions and purchases — was even more difficult to fold into the mix. For subscription-based businesses like ours, the purchase behavior of our customers was essential to providing them with the best possible experience. But the data from our payment solutions was outside of our immediate reach. With Adobe Experience Platform, we could finally break down aislamiento and unite every possible data source, en línea and off, to create an evergreen “snapshot” of every customer.
The first step was to create a space for truly unified customer profiles — using Adobe Experience Platform as the single, mega-repository that would house our full data set. While we had unified data within various separate systems, like Adobe Campaign and Adobe Experience Manager, there had been no catch-all customer profile platform for every individual in our database. Consistent, personal, and relevant experiences all depend on having the full picture. Without holistic profiles, there was no single source of truth to capture an accurate view of each customer and allow for the right experience at the right time. As our designated repository for every single customer profile, Adobe Experience Platform gave us the 360-degree view we’d been looking for.
But having a consolidated data set wasn’t the same thing as having the right data when we needed it. Even with one single repository, data still trickled in slowly, moving through processing and publication before becoming usable. And with every additional data source causing more latency, the unified data repository wasn’t coming close to real time yet. Without getting our profles up to speed, we were spending time and energy creating perfectly targeted campaigns for the customers of two days ago — not today. In order to deliver high-quality experiences, we needed to reevaluate how we gathered our data in the first place.
As Block says, “Organizations like ours are looking at real-time streams and not waiting for latent data to arrive and be processed before being actioned.”
Adobe Experience Platform streaming capabilities, especially for offline events, is what helped us begin to deliver the most relevant experiences possible. While data lakes are still an important part of our customer profiles, streaming offline events added the critical element of purchase and account behavior. Experience Platform uses Apache’s Kafka — implemented in a NoSQL environment using Mongo DB — to stream our offline data. As Block states, this decision helped us bridge the data gap we’d been working to solve.
With online and offline events captured in one place in real time, we now had the big-picture view of our customers that we needed to deliver the caliber of experiences we were aiming for.
For over three years, we tested, tweaked, and perfected Adobe Experience Platform to transform it into a solution unlike any other in the marketplace. What started out as Marketing Cloud Data Platform evolved into the Customer Activity Hub on Adobe Cloud Platform. And in 2019, Adobe Experience Platform, our third release, became available to the public. As a breakthrough solution, Adobe Experience Platform has given us valuable results — and also taught us what it takes to reap the rewards of real-time.
To get the most out of Experience Platform, we needed the right people and processes — not just the technology. For instance, Adobe had already adopted the data-driven operating model (DDOM) prior to implementing Experience Platform. For organizations like ours, it was imperative that we structure our business to prioritize data in a way that supports positive customer experiences. DDOM is one way of getting leadership on board, aligned, and accountable. In this particular model, companies link KPIs to every step of the customer journey so there’s a direct relationship between data and business performance. We didn’t require a new operating model to implement Experience Platform — but already having it in place made implementation that much smoother.
Perhaps the most beneficial precursor to implementing Experience Platform was simply adopting a “data as a product” mentality. In other words, treating data as every bit as business critical as products themselves. While most organizations already understand the value of data, reframing it as a product helped us build out critical operational — and mental — space for it within our organization. As Block said, “Thinking of your data and customer profiles as products really helps crystallize its value proposition. It’s not just a collection of tables or attributes, and it’s not something that ‘somebody else’ will take care of. You have to make it a priority and be ready to learn along the way.” Because when organizations treat data as seriously as their core products, customers reap the benefits.
Implementing a new product — even our own— always requires care and delicacy. We paid special attention to how the rollout would affect our internal teams — especially our marketers, who rely every day on Adobe products to get their job done. Any significant downtime would adversely affect our teams’ ability to do their job. But because Adobe.com already uses Adobe Audience Manager for segmentation and Adobe Target for A/B testing and personalization, it was easy to implement Experience Platform while minimizing disruption. For Adobe’s marketing teams, the Audience Manager profiles they were already used to working with simply refreshed, pulling in attributes from the new unified customer profiles. Though the change may have seemed small at the time, this marked the beginning of the use of Experience Platform in-market.
With consistent, real-time data on every customer, our marketing teams had more tools than ever for creating remarkable experiences. In our initial minimum viable product (MVP) release, we added nearly 200 traits to Audience Manager to add nuance and richness to our understanding of Adobe customers. Among many others, these traits included entitlements, funnel status, user personas, and product usage — with additional traits being added regularly. Plus, instead of the usual 24-hour latencies to update these traits, Experience Platform updated in mere seconds.
And that’s just for starters.
Many years, releases, and milestones later, Adobe Experience Platform has helped us come closer to delivering the quality of customer experiences we want to see in the world. Today, we’re syncing up to 600,000 unified profiles a day for visitors to Adobe.com. Every Adobe customer who visits the website gets a seamless, personalized experience based on their unique, unified profile stemming from over 13 realtime streaming events. And all of this happens whether or not customers are signed into their accounts. This way, we can encourage newer customers to develop their knowledge, give expert customers a deep-dive into their most-used products, and even deploy win-back campaigns for former customers — even when they aren’t logged in. In combination with our other solutions, we can now offer a fully personalized, real-time customer experience across every channel for our 970+ million profiles.
All of which update in 10 to 14 seconds, instead of the 24 to 72 hours it used to take.
Even with all the progress we’ve made, we’re nowhere close to stopping. We’ve broken down the aislamiento that kept us from forming comprehensive customer profiles. But there are still boundaries to break — including predictive modeling for delivering experiences. As we work to trim down the seconds it takes to update profiles, we’re also looking at ways we can meet customer needs before they ever arise. Per Singh, “With data science capabilities that the platform provides, machine learning or artificial intelligence-based models could predict what a customer is going to do next.” The right combination of demographic, en línea, and offline data, with the power of Adobe Sensei, will start giving us a realistic picture of customer intentions. And with every problem we can solve ahead of time, we free up customers to accomplish whatever they’ve set out to do.
With millions of customers empowered to move faster, dig deeper, and push further, every experience adds up. The cumulative effect is significant. And it will help us accomplish what we’ve set out to do — change the world, one digital experience at a time.