At pretty much the same moment I was publishing my post on the merger of martech and adtech into madtech, Adobe was announcing its latest marketing products, including a press release on uniting “Data-driven Marketing and Ad Tech” . Naturally, this caught my attention.
As you might expect, Adobe’s reality is considerably more complicated than the simplicity of the “madtech” vision. Like the other enterprise software vendors who offer broad martech and adtech solutions, Adobe has built its marketing cloud by buying specialist systems. And, again like its competitors, it has only integrated them to a limited degree.
In Adobe’s case, the various products remain as distinct “solutions” served by a common set of “core services”. The current set of nine solutions includes Analytics (Web analytics, née Omniture Site Catalyst), Social (social publishing, based on Context Optional), Target (Web optimization and personalization, derived from Omniture Test & Target/Offermatica), Experience Manager (Web content management , originally Day Software), Media Optimizer (based on Efficient Frontier and Demdex), Campaign (formerly Neolane); Primetime (addressable TV) and Audience Manager (data management platform, formerly Demdex). Of course, the products have all been modified to some degree since their acquisitions. But each still has its own data store, business logic and execution components.
Rather than replacing these components with common systems, Adobe has enabled a certain amount of sharing through its core services. In the case of customer data, the “profiles and audiences” core service maintains a Master ID that is mapped to identities in the different solutions. This means that even though customer data gathered by each solution stays in that solution’s own database, the core service can connect the data in queries to build audience segments. Audiences, which are lists of Master IDs, can also be defined in one solution, sent to the core service, and then shared with other solutions. Data from external systems can be imported to the core service in batch processes, stored within the core service’s own database, and also used in segmentation.
Adobe says that data stored in the solutions can be accessed in real time. I haven’t had a chance to assess how well this works in practice, although my default assumption is that a shared central data store will work better than access to distributed data. But there’s a case to be made that digital volumes are so huge and change so quickly that it would be impractical to copy data from the solutions to a central database. So maybe Adobe’s approach is best. Certainly where external data is concerned, marketers will increasingly have no choice but to rely on distributed data access.
Still, distributed systems often don’t keep the kinds of aggregated historical data, such as trends in model scores, that are needed for marketing purposes. So I’d expect that marketers will want at least some persistent data of their own even when they are also reading a lot of external data. Nevertheless, Adobe’s approach of storing on Member IDs and lists of Member IDs in the shared cores services does provide a viable approach to building a complete, omni-channel customer view.
But here’s the catch: Adobe’s approach only works if all your systems are actually tied into the central system. Adobe recognizes this and is working on it, but so far has only integrated four of its nine solutions with the profiles and audiences core service. These are Analytics, Target, Campaign and Audience Manager. The balance will be added over time.
The second big limit to Adobe’s current approach is sharing with external systems. Only Adobe solutions can access other solutions’ data through core services. This makes it difficult to substitute an external product if you already have one in place for a particular function or don’t like Adobe’s solution.
Adobe does connect with non-Adobe systems through Audience Manager, its data management platform, which can exchange data with a company’s own CRM or operational databases, business partners, and external data pools and ad networks. Audience Manager can hold vast amounts of detailed data, but does not store personally identifiable information such as names or email addresses. Audience Manager can also copy Web behavior information directly from Analytics, the one instance (so far as I know) where detailed data is shared between Adobe solutions.
Confused yet? Well, wrap your head around this one: Media Optimizer, Adobe’s online ad manager (a.k.a. demand side platform or DSP), isn’t yet integrated with core services but can also import audience segments from Audience Manager and Adobe Analytics. It will soon be able to import segments created in Campaign and Target as well.
So far, I’ve only been discussing data integration. The various Adobe components also have their own tools for segmentation, decision logic, content creation, and other functions. These are also slowly converging across products: for example, there is an “assets” core service that provides a central asset library whose components can be uploaded to at least some of the individual solutions. The segmentation interface is also being standardized product-by-product. There’s no point in trying to list exactly what is and isn’t standard today, since this will only change over time.
The lesson here is that suites are not simple. Marketers considering Adobe or any other Marketing Cloud need to examine the details of the architectures, integration, and consistency among the components they plan on using. The differences can be subtle and the vendors often don’t explain them very clearly. But it pays to dig in: the answers have a big impact on whether the system you choose will deliver the results you expect.
