The Anonymous Problem for Marketers

Across the common marketing stack, marketing teams are using 10+ tools to effectively personalise experiences for roughly 2-3% of their customers and prospects. (ie. A prospect or known customer where an email address, phone number, app install or postal address has been collected.) However, targeting and personalising experiences for the 97-98% of anonymous prospects across digital channels, who don’t authenticate, is an industry-wide issue that the majority of MarTech solutions don’t want to talk about.

?

It obviously makes sense for marketing automation (email, mobile and other channel) solutions because you can’t email someone without their email address, you can’t send a mobile message without the phone number etc. The result is that these platforms are investing in zero-party data capture tools to try and convert more of the 97-98% of anonymous site visitors into handing over their contact details, enabling these tools to do what they do best.

?

However, with a huge volume of a brands’ prospects choosing not to convert (regardless of attempts to do so), there is a huge potential that very few technologies are focussed on fixing, and while Customer Data Platforms (CDPs) claim to be a central point for all customer data, they too have not done enough to help drive better acquisition results from real-time anonymous prospect data. Remember, that's 97-98% of your funnel!

?

The Anonymous Activation Gap: An Ongoing Challenge

Movements in the MarTech world, particularly in the CDP market, mean the problem is about to get worse still! The CDP world has essentially been going through a period of rapid change as the technology has begun to fall into 2 distinct buckets – Profile Builder CDPs (trending towards a warehouse-native approach to data storage) and Campaign CDPs (trending towards marketing automation capabilities and non-warehouse native). ( David Raab, the founder of the CDP Institute, has an excellent recent article on this here.)

?

Where businesses are opting to utilise their cloud Data Warehouse as the central store for all customer data, the reality is that putting anonymous clickstream data into the data warehouse is going to create significant increases in data processing cost, not to mention the struggles to satisfy true real-time use cases. With ‘Profile builder’ CDPs leaning towards this trend, Tealium’s recent Snowflake integration (based on their press release) seems to offer ‘warehouse-native’ composability but with the capacity to run the real-time use cases on the fly that you’d expect from a traditional CDP. However, this type of solution is geared towards enterprise brands with deep pockets to invest in a CDP, a Cloud Data Warehouse AND a range of ‘best of breed’ execution platforms.?

?

This approach, while expensive, provides all the capabilities to minimise data storage for known customers with a warehouse-native approach to customer marketing, while also providing real-time anonymous capabilities via the CDP for unknown prospects.

?

In general, we are seeing all the traditional ‘packaged CDPs’ (more suited to the large enterprise customers) move ever closer to the data warehouse, ironically making them very similar to composable warehouse native solutions that have become more popular over the recent years. They might not be at the Tealium/Snowflake level but the latest release of new capabilities for real-time personalisation and/or journey orchestration are headed that way.?

?

But who therefore is going to fix the anonymous problem for brands that cannot invest large sums of money in enterprise solutions and cloud data warehouse projects?

?

Why Composable CDPs were never the answer to the anonymous problem…

The composable CDP was expected to replace the traditional CDP (Hightouch has grown rapidly pushing a zero-copy approach to CDPs), but instead, they have forced innovation from these players (see Treasure Data, MParticle and Tealium’s ?recent developments). As the connectors from a traditional CDP and a data warehouse become more complex and varied, it becomes harder to distinguish between the two.?

?

Of course, there is the concept of ‘zero copy’ data, but this is where the difficulties begin. If everything is processed in real-time in the data warehouse, then it becomes very, very costly and a nonstarter. So, any solution that relies solely on the data warehouse cannot run use cases around anonymous visitors or real-time. If you use a traditional CDP on top, you have to start copying data across the two platforms - you can’t have your cake and eat it.?

?

The fully composable solutions have therefore started to build out offerings to fix this issue, but as a result must start to also clone features and functionality (and caching of data) that traditional CDPs already do. Therefore, the distinction between composable and non-composable CDPs has blurred, and without innovation towards fixing the anonymous problem, the composable CDPs will only ever provide a strong emphasis on known customers and offer very little for converting prospects to customers. Without the ability to copy any data and act on it the composable solutions become a slave to the processing of the data warehouse.

?

Implications on the use of AI

The answer then would be to accept that you cannot have one platform and buy a CDP to sit on top of the warehouse, and minimise data ingestion into . This sounds great, but ultimately this misses out on the single biggest? innovation in the industry - AI!

?

For AI to work at its best, it needs to be across all the data of an enterprise and making decisions across all touch points. How can you do this if the data warehouse has some of the data, and the CDP has another chunk of the data? You might be able to reconcile it at some point, but is that going to be in real-time? The recent Tealium/Snowflake integration seems to try and bridge some of that gap, but only time will tell if the two organisations are genuinely working as closely as claimed.

?

So, perhaps there is room for smaller CDP offerings that focus solely on acquisition? Using AI to personalise the onsite/app experience, and then retargeting those visitors via social and Google? This would reduce the cost significantly as no customer data would need to be stored and would fill a genuine gap.?

?

How can non-enterprise brands leverage anonymous profiles?

For the non-enterprise companies there has also been a significant amount of change. The CDP market was perhaps too expensive at first, largely because companies were trying to orchestrate activity across multiple best-in-breed platforms. What we have seen over the last few years is the growth of marketing automation tools that offer a one-stop solution for personalisation, email, SMS and app-based marketing to known customers

?

Whilst these platforms aren’t always best of breed in every channel, they can provide simplicity and deliver results. It’s far easier to orchestrate customer journeys from one platform that includes execution. As these platforms have become stronger they have even come to label themselves as CDPs, to the point that almost anything is a CDP these days!

?

However gaps remain here for the anonymous visitor too. If we are to call these platforms CDPs, then the best way to characterise them (as David Raab says) is as a ‘campaign CDP’ that generally starts with an email (and therefore a known user). The ID resolution is not as robust as an enterprise CDP, and will not stitch anonymous profiles together across separate visits, so this again becomes a customer focussed tool and offers little to convert prospects to customers.

?

The anonymous offerings they have are generally not built to stitch together multi-visit anonymous IDs, but rather clickstream in-session browsing (at best). They can sit on top of a Data Warehouse to try and use previous transactions and behaviors from the known customers, but this connection is not enterprise level (and nothing close to what Tealium proposes with Snowflake in that space). Added to the fact they are not real-time, and again, we see a limitation on use cases.?

?

As we see the CDP market change I think it’s inevitable that the marketing automation platforms will either build out more CDP functionality to address the anonymous identity issue. That begins with ID resolution and deeper profile building, whereby the marketing automation platforms will either build or integrate with CDP offerings that can bridge that gap.?

?

In my opinion large enterprises will either look for a stand-alone solution for handling anonymous acquisition use cases, and run that solution alongside a composable CDP for known customers. We will likely see the abandonment of the hype of ‘zero copy’ and instead see an era of ‘minimal copy’ as traditional CDPs innovate to offer better data warehouse integration, much like the new Tealium/Snowflake partnership. This will see the CDP use only the data required for anonymous activation and real-time use cases, while the known customers will be activated from a data warehouse.?

?

And for companies without the budgets for a CDP, a cloud data warehouse AND stand-alone execution platforms will be catered for by the ever improving marketing automation tools. I expect to see these tools incorporate real-time, genuine ID resolution and better data warehouse integrations too.

?

This will provide businesses of all sizes with the ability to market to both anonymous visitors AND known customers from one platform, whereby they will be able to convert prospects to customers and customers to loyal customers.

?

So, how are you fixing the anonymous problem?