We need to stop talking about silos. I realise the irony of that statement being placed at the beginning of an article about silos. A silo is a dry term used in media to define individual departments that don’t connect. It’s also something that as digital marketing has developed at a pace and despite the opportunities on offer to connect, has become more frequent.
Marketing is a department blessed with variety. A mixture of creative and functional tasks that help generate sales and brand awareness all come together to create the voice of the company. With those activities come a range of performance points – some firm data points, some less concrete. Rarely do any of the metrics meet up in the middle. Multi-touch attribution methods work to address some of these, but they’re not without their flaws.
While there are still silos within marketing departments, there will be attribution issues. Channel buying is done in silos all over marketing landscapes – from in-house buying to agency planning.
To clarify what I mean by channels (as there are many misclassifications out there) – mobile is not a channel; it is a device, just as a desktop is a device and video is a format. However, as users behave differently with mobiles, it needs to be planned differently. Search, on the other hand, is a channel as you can search across different devices. As planning is often done in silos, display and mobile are often fighting against each other, when actually mobile should be incorporated into display – after all, it’s the consumer we should be focusing on, rather than a channel. It also means often there is an overlapping of effort.
While it may be easy for marketers to plan activity by portioning out spend by channel, this creates an attribution problem when data from multiple platforms is collected and analysed separately. What one company may use to identify a customer will likely differ from another company’s identifier. This makes it difficult to recognise individuals across multiple platforms, and even more difficult to measure. Compounding the problem is that suppliers are creating walled gardens to protect their data, creating silos within silos making measurement even harder.
Currently, cross device measurement and targeting is not working fully for anyone. Each DSP, data and technology provider in the ecosystem has different rules around layering data, which companies can see it and connect with what. For example, some media owners stop advertisers from using some tracking technologies within their ecosystem. While advances are being made in cross device abilities, siloed approaches to tracking are actually getting worse, and as walled gardens continue to go up, how can marketers glue it all together?
Just as last-click attribution does not tell the story of the consumer’s path to purchase online, looking at online marketing channels alone can overstate the effectiveness of them, leading to bias. Econometrics models which incorporate multiple interactions online and offline are the only way true attribution can be achieved. However, given the changing nature of consumer habits that also develop over time, such as social’s growing importance in word-of-mouth marketing, any model also needs to incorporate this into their calculations.
There is a strong argument to say that in the future, as all channels become more digitally enabled, they will naturally become closer together. TV, OOH and radio are all moving digital. However, it doesn’t explain why digital channels, such as search and programmatic display are still planned separately now, despite there being many similarities.
Connected consumers will create connected data. When the promise of truly integrated data is fulfilled, the silos will erode and we’ll have a centralised data centre. Until then, we’ll have to put up with more silo chatter and broken attribution methods.