Thought Leadership April 17th 2020

Capturing the explosion in demand: keeping customers in new habits through marketing optimization

The COVID-19 pandemic has sparked a global economic downturn greater than anything seen since the 2008 financial crisis, but for certain businesses it has driven an exceptional boost in demand.

Meal delivery services, streaming services, wine clubs, online grocery, and e-commerce arms of companies are all benefiting from stay-at-home orders and the pausing of physical selling during this period. UK e-comm grocery retailer Ocado has seen Google search increases of 550%, increases at the recipe kit delivery service Hello Fresh have hit 260% growth, and searches for Peloton bikes have hit levels only normally seen in the run up to Christmas[1].

These businesses no longer need to invest in marketing in this landscape; their demand far outstrips their ability to service it. However, the real challenge (beyond the short-term operational capacity) is how to ensure the serviced demand now is not lost as social restrictions are eased and consumers are able to shop more as they used to. If captured effectively, this can be the single biggest growth driver for many businesses that they will ever have; the question is, how?


Identifying the potential loyal customers

All customers that have signed on to new services and methods of delivery through the crisis are likely doing so as it has become more difficult to buy through their habitual routes. What is absolutely fundamental to successfully capturing this increase in demand long-term is to separate those that are more likely to stick to this new habit if engaged in the right way vs. those that will quickly transition back. The data available through e-commerce purchases, giving direct knowledge of the customer and a route to contacting them (with the right permissions, of course), means that the businesses experiencing surges in demand can already be planning who and how to target.

There are many features of customers and their purchase behavior that can point to the likelihood of a customer becoming loyal or high value in the future. What is very clear is that simple models in the past based on basket size are no longer relevant through the crisis. Instead a much more nuanced approach needs to be taken to identify your potential loyal customers.

The first step is to truly understand the differentiating drivers of loyalty and value from the ground truth that is the past. Proximity to brick & mortar competitors, estimated size of household, geographic location, demographics; there are an enormous number of features that can be constructed from captured data to help explain what distinguishes those who are loyal from those who are fleeting. Some of them are directly owned by brands, collected through the normal sales process; others need to be constructed by bringing together the right external sources (eg. maps data, for getting competitor locations) and internal data (customer address to measure proximity).

Once this more detailed model of loyalty is constructed, one that is more resilient to the changes in demand through COVID-19, the discriminating features can be discovered amongst the surge of new customers to identify who deserves your attention the most.  This enables a shift in marketing from pure acquisition in marketing levers and retention through CRM, to a marketing organization completely focused on keeping the step change into whatever our ‘new normal’ economy looks like.

For each industry, the exact approach will be different. In grocery, large basket sizes and high frequency are no longer a valid indicator; these just reflect the changing consumer need during this period. Homes are fuller throughout the week and so more food is needed to sustain them. However, other indicators, such as the type of product bought can still be used. Looking at customers that purchase products priced above the category average can be a good indicator of potential value (but there is a need to be careful with stock availability distorting results here). Alternatively, a high number of meals chosen in a recipe delivery service is likely due to difficulty in accessing groceries now, but enhancing address data with external sources on property size and purchase price can be an indicator as to whether the household makeup is a good fit vs. those that have been loyal in the past.


Making new habits stick

The other side to the coin is how to activate marketing to retain your high value customers. Direct communications through CRM initiatives are clearly a prime source for this, and the identification of your most likely customers to be high value lets you focus more costly CRM efforts on those that deserve it the most. However, contactability and attention from those you can contact can be an issue. Low opt-in, open and click-through rates mean that you need to ensure broader marketing activities are also playing their role in retention.

Most of the time, CRM strategy and Marketing Mix Optimisation have been kept separate in the past; perhaps with CRM actions being included in an econometrics model to understand their impact. However, there is an opportunity to go much further and approach these as a linked discipline, increasing the strategic value and impact of each. This powerful new capability will ultimately allow you to steer your investments between retention focused and acquisition focused as the business and your tracked KPIs evolve.

By understanding which levers across the whole marketing mix are able to drive repeat visits from high value customers, businesses can put the plans in place to retain as much of the growth they’ve seen in the last weeks as possible. Marketing Mix Optimization can play a crucial role here and ensure you’re not just steering from the limited perspective of an attribution model. Constructing econometric models both for customer acquisition, and for retention of high value customers can allow you to view which levers over-index in which role – acquiring new business or keeping the most valuable, as well as how these interact with external drivers.

Some levers will be equally strong at both and can be leant on to continue new acquisition as well as playing a role in retaining a customer base. However, others will skew much more strongly to one role vs. another. Immediately post-COVID-19, as social restrictions are lifted, for many businesses the priority should be in retention. This approach allows those businesses to prepare a marketing mix that captures the spike in demand they have seen in the short-term as a step-change in business performance. This also leaves them with the clarity of how to transition to a more balanced plan, and to flex throughout the year as KPI tracking of acquisition vs. retention fluctuates.


What next?

In short:

  1. Serve your customers to the best of your ability during the demand spikes

  2. Ensure you are able to identify customers that are the most likely to become loyal and keep new habits

  3. Drive the best communication for all these with targeted actions and a marketing mix shifted to high value retention

Some teams right now are rightly focusing on servicing operational demand beyond anything they have seen before. It is critical to provide as high a quality of service as possible to demonstrate the benefits of your model during this period and leave customers with a positive impression. However, those businesses that are then planning on how to retain this new customer base, and are commissioning and running the analytics to chart the direction, will be best placed to see long-term growth and avoid customers moving back to their old purchasing habits. 

[1] Google Trends: UK Market


By Matt Andrew, Managing Director - Ekimetrics London


Contact us for more information:


If you would like to know more why we believe data-driven approaches are key to bringing agility and marketing intelligence to strategic decision-making, register now to the coming Data Science webinar!


News May 6th 2020

Alcov2: a large-scale survey to study the transmission of SARS-CoV2 within French households

A team of researchers with expertise in mathematics, statistics and epidemiology (Sorbonne University, CNRS, Collège de France, Oxford University, Ekimetrics and Datacraft) has launched an investigation into the French outbreaks of Coronavirus during the recently implemented containment period.