The definition of real time in the context of marketing varies by use case. For a retailer implementing a buy online, pick up in-store (BOPIS) model, real time could mean engaging with a customer with a personalized message or offer within five minutes of the online transaction. For a web services company, real time will likely require a much faster interaction, likely within milliseconds of a customer appearing in a channel of engagement.
Determining the meaning of real time depends on what a marketing organization ultimately wishes to accomplish by eliminating latency between the ingestion of data from any source, applying analytics, and taking action. Business goals must be weighed against operational and cost considerations, and measuring those considerations on a sliding scale against potential lift from the introduction of real time. Will making a process faster generate revenue that will offset the additional cost and resources?
Real Time and the Next-Best-Action
While real time can transform many marketing strategies, including onboarding, retention, and acquisition, it’s important to understand that real-time data is only one of three pillars that make up true, transformational real-time capabilities. Real-time access to every customer data source, combined with real-time decisioning through in-line analytical models and real-time action – the ability to enact the decision at the precise moment of interaction – together make up the core capabilities of real-time marketing.
For a web services company to know everything there is to know about a customer exactly when that customer engages with a chatbot is of minimal benefit unless that recognition generates insight for a next-best action for the marketer to take during the interaction. For the BOPIS retailer that recognizes an in-store customer as the same person who just purchased a chainsaw online, that real-time action could be a discount offer for an accessory that was placed in an abandoned shopping cart.
Real time in the context of generating a next-best offer does not mean that a brand has a static offer queued up ready to go regardless of where a customer next appears in their buying journey. An important distinction is that the next-best action is generated based on the specific interaction or channel. Machine learning models that drive intelligent orchestration and real-time decisioning start firing on all cylinders and make a calculus according to a precise moment of the customer journey.
In the BOPIS example, a next-best action may differ based on the customer’s behavior when picking up the purchased item. A discount offer for a chainsaw cleaning kit might be proffered if the customer goes right to check-out, but if the customer opens the mobile app while lingering in the home improvement aisle, the next-best action might be a personalized direct mailing with kitchen remodel ideas.
Real Time Quick Wins
Factoring in revenue lift that can be generated by introducing real time into a marketing strategy can be the deciding factor on whether real time for a business means five milliseconds, five minutes, somewhere in between or longer. On-boarding is a particularly high-value journey that stands to benefit from real time. Introducing real time into a strategy for welcoming a new customer, introducing them to the brand, and encouraging the customer to follow the brand on social media can have a significant impact on customer lifetime value (CLV), in line with the old “no second chance to make a good first impression” adage. A brand that asks a customer to follow them on Facebook when the customer is already a follower will likely introduce friction into the onboarding experience.
Armed with a real-time single customer view, a brand can engage with each new customer in the context and cadence of an individual customer journey, offering a personalized experience that recognizes the fact that customers onboard at different paces.
Customer renewal strategy is another area where real time can deliver big wins. Marketers know that it’s less expensive to keep an existing customer than acquire a new one, and real time allows a marketer to throttle retention engagement according to a customer’s unique behaviors. A lapsing customer, for example, might be met with a more enticing retention offer during interactions than might a frequent transactor who would convert with a different message.
Real-time capabilities often significantly lower the cost per acquisition/retention of a customer. Continual testing of models can help place a monetary value on real time. What percentage of customers did you retain with a same-day email offer versus a second-day email? Does conversion rate improve on your website when content is presented based on propensity scores and clickstream data?
A decision of whether to implement real-time capabilities must include real-world considerations. Marketing must be wary of overpromising. A customer service chat complaint that an item wasn’t delivered in the timeframe promised might generate a next-best action of an apology with refunded shipping costs, but marshaling additional resources – re-routing trucks, white-glove service, etc. – might be out of scope. Marketers must balance a fine line between being responsive enough without adding unnecessary cost or resources.
Cost considerations are not the only consideration when making the determination for how and whether to implement a real-time marketing strategy. A millisecond response time might be ideal for a web services company but introduce the dreaded “creep factor” for a retailer.
Real Time Runs on Metadata
A brand with a single customer view; a golden record for each customer, will have at their fingertips everything there is to know about a customer – behaviors, transactions, preferences. Real-time marketers score their customer file for various outcomes such as persona, segmentation, LTV score among many others.
Delivering a real-time engagement in the context and cadence of a customer journey, machine learning algorithms determine next-best-action or recommended products based on metadata as well as event data. Did the customer buy or not buy an item? Did they fulfill or abandon a shopping cart? A trigger or interaction that generates a real-time response may only need yesterday’s data, for instance. Not every part of a golden customer record is going to be live, event data.
When creating new models, the decision about what data is needed can be as important as deciding how fast a real-time engagement needs to be. For a real-time “before you go” message for a customer who abandons a shopping cart, will mining 10 years of transactions create a more personalized engagement than looking at just the past 30 days? If so, does that outweigh the extra millisecond, second, or minute that churning through more data will take?
Real time is about collapsing the time between data, decisions, and interactions in an outcome-focused way. Measuring outcomes such as revenue lift or customer satisfaction will help marketers make important decisions from a time, cost, and resource perspective about what real time means in accordance with their business goals.