The Return of Data-Driven Marketing

It's event season in the MarTech world. Over the past month we have had presentations from the leading analysts and technology vendors about the state of the market and the direction in which the industry is heading. I've written about both Oracle Modern CX and Adobe Summit over the past few weeks. There was a fair degree of commonality between the announcements by both firms at their annual summits even though they are pursuing different strategies.

Data-driven marketing is back in the conversation, due to a renewed emphasis on single customer view and AI enablement. If this sounds familiar, that's because we're retreading old ground. B2B marketers have been here before. In the early days of marketing automation, deployments were often justified by the lack of any central marketing database. Many businesses had tried to use CRM systems for that purpose, but quickly discovered that the conflicting needs of Sales and Marketing departments meant that a new solution was required.

Early Promises

Enter marketing automation with its native lead scoring and web tracking capabilities, as well as a direct line into the all important CRM system for Sales follow-up. The revolutionary capacity offered by this brand new technology was the possibility of mixing engagement history data and contact profile data when building audience segments or making decisions in workflows. Combined with the native reporting tools, it became possible to execute outbound email campaigns using data from previous campaigns to choose the right message, select the right audience and deliver at the right time.

In practice, things haven't worked out quite as expected. Reporting has always been a major pain point for marketing automation users, and the web tracking capabilities of Eloqua and Marketo are extremely basic. This makes it challenging to integrate web activity into campaign segmentation within those platforms because the only activity options are a list of page views and form submissions. None of this data has context and ignores the wide range of web interactions that aren’t page views such as interactive content clicks and video plays. On top of this, there is an extensive portfolio of digital activity run offsite to be considered. There are workarounds for all these things, but they require either manual effort or custom development. Even more criminally, the reporting and tracking limitations also apply to the core email capability of marketing automation. In some systems, it can be difficult to report on which links have been clicked if a particular URL appears twice in an email.

The Data Gap

Every marketer talks about data-driven marketing, but the gaps in tracking digital activity have made it difficult to implement in practice. There's still a lot of data available for Marketers to interpret despite these gaps, but it is siloed across multiple tools and is highly fragmented. The most forward-thinking teams have implemented deep integrations between their MarTech stack and BI tools in order to get as much data as possible for analysts to delve into and interpret. Reporting systems such as Domo and Tableau have seen broad adoption within marketing departments to visualise the stream of analytics and activity data flowing from websites, marketing automation and other systems. Even with these integrations in place, it is still necessary for marketers to join all the various data sources together to build a complete picture and then spot the trends amidst the noise. The sheer volume of data required for this means that marketers need to ask the right questions ahead of time, as there is a long lead time if a different analysis is needed.

Vendors are touting AI as the solution to this problem. AI and its precursors have been around for as long as data-driven marketing has been discussed. Its primary use case so far has been in lead scoring, which is the most widely understood manifestation of the need for data-driven decision making within the marketing organisation. Inaccurate lead scoring results in the wrong leads going to Sales, and the unwelcome high-level executive attention that this often generates. It has long been seen as self-evident that a lead scoring model which uses a hundred data points crunched by data scientists is better than one which uses half a dozen data points crunched by marketing operations. This doesn't necessarily work on practice, but multiple vendors have built very successful businesses on top of this fact, and have developed AI algorithms to aid them.

The Latest Must-Have

Now these same AIs are being retrained for the latest must-have marketing technology – customer data platforms (CDPs). CDPs solve a clear need – the data models of marketing automation platforms are too simple for modern marketing departments, and the data structures of CRM platforms are too focused on Sales needs. Marketing deals with anonymous engagements tracked through a DMP, digital engagements tracked through a MAP and face to face engagements tracked through a CRM, yet lacks a mechanism to tie all these things together into a single database that can then be sliced and diced for decision making, be they funnel related decisions or campaign related decisions. CDPs are designed to do precisely that. They ingest the engagement and profile data from all your other platforms, and then send back lists of contacts or accounts to be targeted for campaigns.

Oracle and Adobe are re-architecting their marketing clouds around their recently launched CDP products, and Salesforce are promising to do the same. They will become the underlying data layer upon which all the other tools in the stack are built. In future, Eloqua and Marketo will pull lists from these new data platforms when selecting audiences or triggering campaigns. The end result is that marketers will finally be able to use prospect activities that occur on one channel such as a content marketing platform or mobile app to trigger on different channels, such as outbound campaigns through marketing automation or digital advertising through a DMP. The CDP will act as the central database of the marketing department bringing together all the data and platforms needed to achieve this into one place.

CDPs are more than just databases or dashboards. A standard database optimised for working with the largest of datasets can't help with organising and interpreting all the information received from the rest of the stack. This limitation starts with the basics such as merging duplicates and allocating contacts to accounts, and extends all the way to making decisions about campaign audiences or email send times. This is where CDPs and the AI buzz attached to them come into play. Your job is to collect all your data in the CDP and tell the platform how the various elements fit together, it then builds a unified account and contact view that can be analysed using reporting tools or pushed into a cross-channel campaign stream. If this sounds like the original vision of lead nurturing and marketing automation, that's because it is.

A Lost Vision

Data-driven marketing was revolutionary not just because it opened up new channels of communication, but also because it made it easier to measure the outcomes of those communications. Marketers could then optimise their audience, timing and messaging based on actual results of previous campaigns. This has happened to a degree, but the avalanche of inaccurate or irrelevant data has made it more difficult than it should be. Marketers have then fallen back on gut instinct to bridge the data gap with predictable results.

New technology is being deployed to bring order to that chaos, so that the initial promise of digital marketing can finally be fulfilled. The pay-off will be in the results you deliver to the business; the trade-off is the additional technology complexity this brings. CDPs have been around in B2C for a while. Adapting them to B2B is a new idea with the potential to bridge the performance gap between the weakest and strongest marketing teams. The road to better campaign results is in the data. You just need to look for it.