There’s an old saying about failing to plan is the same as planning to fail. There are many things that can potentially go sideways and you need to account for as many as you can.
The same rule of thumb applies to retailers building a customer data platform (CDP) who need to be aware of a host of pitfalls that can derail implementations designed to deliver real-time, omnichannel personalized retail experiences.
Here then, without further ado, are the Top Five things retailers need to watch out for when building a CDP to make sure things don’t go horribly wrong.
1. Give Data Quality its Due
One thing that will get things off track in a hurry is if data quality is not given its due as soon as data is ingested from every conceivable source. Retailers should steer clear of a CDP that outsources data quality to a third party, including the cleansing, normalizing and matching of data. This is big. A CDP that outsources or puts off data quality will make it virtually impossible for marketers to deliver an omnichannel customer experience (CX) because they will not trust the accuracy or completeness of customer data. Also, if data quality is completed downstream, the resulting waiting game means there will not be any real-time visibility into how a customer is moving through a customer journey. As every retailer knows, real time is vital for delivering a relevant next-best action the precise moment a customer shows up in a channel.
2. Ongoing Identity Resolution
Conversely, another pitfall retailers need to be wary is of a CDP that stops at basic, deterministic matching. Linking a third-party cookie to a device’s browsing history might suffice for basic personalization, but if a basic match is all you’re doing you are going to lose out on real-time data integration, being able to effectively integrate all types of data, connect to all sources or build aggregates. Some CDP’s, in other words, treat identity resolution as a one-time process, but a point-in-time match does little to reveal anything important about how a customer journey unfolds. Without such a longitudinal view, retailers lack the context they need to be relevant across all channels, at any point in an omnichannel customer journey.
3. A Rules-Based Approach
Retailers also need to be cautious about a CDP that uses a list-based approach to database extractions. In a list-based approach to building an audience, once the list is created it is subject to decay. New data is by definition excluded, meaning a campaign is built around certain attributes of an audience that may no longer exist. Redpoint rg1, by contrast, adopts a rules-based approach where an audience is created at the last possible moment before any inflection point, according to dynamic rules in place – a set of logic that is evaluated at any point in a campaign where a list would typically be used. A campaign can then be designed for a living, breathing audience that is relevant to an individual at a precise moment of a customer journey.
Some CDP’s treat identity resolution as a one-time process, but a point-in-time match does little to reveal anything important about how a customer journey unfolds.
Consider a direct mail campaign as one example. In a list-based approach, a retailer may send a flyer or coupon to an audience that has shown an affinity for blue quarter-zip pullovers based on browsing history or shopping cart behavior. Once the list is extracted, that mailer will go out to everyone on it – including anyone in the audience who has purchased the pullover in the interim, who has filled a shopping cart with different items, or who has changed a physical or email address and thus will not receive the otherwise relevant piece of content.
With rg1, by contrast, the rules-based approach to database extractions excludes all those possibilities, ensuring a direct mail piece is hyper-relevant to the constitution of the audience at the last possible moment before the mailer is sent.
4. All Customer Data, All Sources
Another snag for retailers interested in delivering an omnichannel CX is a CDP that limits data integration to just two or three channels or sources of customer data. For some retailers, particularly those with channel-specific marketing teams (email, website, etc.), connecting even two channels is a dream scenario. Again, though, as we’ve seen with some of the other pitfalls, while two channels is better than one, it still falls fall short of meeting expectations for an omnichannel CX. Integrating customer data from the website and email doesn’t help a call center agent fielding a call from a customer trying to resolve a returns problem, for example.
Integrating data from every conceivable source of customer data and every type (structured, unstructured, batch, etc.) is another strength of rg1. Because every piece of customer data may reveal an important signal about a customer’s behavior, preference or intent, a key to being hyper-relevant is to ensure the accuracy and completeness of a golden customer record, or Customer 360. Ignoring customer data – any customer data – has the potential to introduce irrelevance into any next-best action.
5. Machine Learning
Lastly, there is also a lot that can go wrong when an organization discounts the use of AI or machine learning in delivering an omnichannel CX at scale. Marketing teams accustomed to generalized segmentation often do not appreciate a need to introduce machine learning; segments are broad enough to accommodate manual changes – particularly if a campaign does not aspire to (or is not capable of) a real-time component.
Of course, as with list-based vs. a rules-based approach, machine learning is instrumental in accounting for a dynamically changing audience. Machine learning built into rg1 provides marketers with hands-off intelligence that automates the continual creation of new, granular segments. In-line, self-training models optimized for a certain outcome (retention, acquisition, lifetime value, etc.) will produce a next-best action for an audience of one precisely tuned for a specific moment in the customer journey and optimized by channel, all without manual intervention.
To recap, while there might be many things that can go wrong when building a CDP for retail, there is also a lot that can go right when a CDP takes care of data quality at ingest, treats identity resolution as an ongoing process, adopts a rules-based approach to database extractions, connects all sources of customer data and, finally, has built-in machine learning.
Retailers can build a very effective CDP when implementing Redpoint’s rg1 for producing consistently relevant, hyper-personalized customer experiences on any channel in any retail environment.
For more on how you can meet every retail customer with highly personalized experiences in every interaction and in every channel, click here.
Be in-the-know with all the latest customer engagement, data management, and Redpoint Global news by following us on LinkedIn, Twitter, and Facebook.
