Guest comment: Direct-to-consumer commerce is data-driven

Direct-to-consumer commerce is data-driven

Easy to start but hard to scale is the reality that has bitten direct-to-consumer (DTC) brands in the last two years. David Hawkings SVP EMEA, antuit.ai says that there is an answer for both pure DTC brands and traditional brands developing a DTC business, and it lies in both the data and the engine that drives it.

Since the pandemic, the excitement over the growth of a slew of so-called disruptive brands that used a subscription-based model to go direct via the Internet, has cooled.

Now in 2021, DTC brands, both the new and disruptive, in addition to the well-established, are rethinking distribution and retail. There has been a refocus on physical retail stores. Some brands have withdrawn from channels, such as marketplaces that were judged to be too expensive and gave brands little access to what should be their own data. While other brands have recognised that their future lies in working with the very marketplaces they once sought to avoid.

So, what is the way forward for DTC brands? Data. Not only do they need to own their data, but they need to be able to exploit it intelligently in order to reduce the increasingly unbearable costs of acquisition and third-party distribution that are currently eating into costs and not delivering conversion.

They then need to be able to optimise customer reach and acquisition across all of their channels, which creates enormous complexity, exacerbated by the fact that new channels are emerging while others are delivering overnight value. Each of these channels comes with their own data and distribution problems, not helped by the fact that during and post pandemic, DTC brands are having to cope with the difficulties of smaller head offices and people working remotely.

This logic then applies for traditional brands that either lack, or are starting to build, DTC capability in order to keep and maintain market share. The proliferation of channels requires both types of businesses to rethink their data and how they exploit it, as well as having to cope with the fact that consumers are purchasing more online than before the pandemic. This is a trend that is expected to continue post-pandemic with the associated risk that brands lose touch with customers, forcing them to rely even more heavily on data.

Currently, data not only lacks depth in terms of customer preferences and cross-channel activity, but there is limited integration with third party sources such as social media. Moreover, allocation, distribution and marketing are often managed in silos, channel by channel. Even within a channel, there is heavy reliance on historical data, a bedrock that has been crushed by customer behaviour since the start of the pandemic. And finally, this limited, single channel view of data is then often managed using spreadsheets which makes forecasts inaccurate and certainly not reliable.

And all this presupposes that the available data is well understood, given that brands going direct tend to not have the experience of working with the new types of data that digital commerce creates including clicks, ratings, buy boxes, etc.

Our recommendation is that brands create a unified demand forecast that will give them complete visibility across stock, channels, distribution, and activity which will immediately provide visibility to relative cost and profitability of each channel, along with the various interdependencies, in order to plan better for the future.

We also recommend that artificial intelligence be deployed to cut the manual effort to produce the unified demand forecast and surface insights that simply cannot be done with conventional analysis, insights that will be critical to providing a competitive advantage. AI can also be used to help cleanse, “fill in the gaps”, and make sense of the new digital commerce data, which is often a daunting task for a Business/IT team, and will be an important starting point to integrating data from different departments, functions and external sources.

The process of getting started does contain several minefields that can cause delays or lead to a less valuable outcome. The first is the debate over buy versus build, which is to some extent a distraction given the fact that both can work side by side, particularly using a third-party solution that can support in-house data scientists. The second is the solution itself. Many tools on the market sell based on the interfaces more than the actual solution capabilities, masking the fact that the engine can only be tuned in a limited way, if at all. They often can deliver some value quickly, but it is not sustainable nor scalable.

The best option is to find a solution that is powerful enough to leverage and manage data from today’s world, as well as be tailored to any scenario that a brand may have.

By David Hawkings
SVP EMEA
antuit.ai