Forecasting, when the past is no longer as reliable a predictor of the future...
As we emerge from lockdown and hopefully embrace a period of recovery and reopening, we cannot ignore the profound impact the last year has had on how we live, and with that how we consume, transact, and engage with the world around us. How then, can brands effectively plan for what is next?
This was the subject of a recent round table debate with some of the UKs leading econometricians. During which, the panel comprising Steve Messenger, Clive Frostick and Bryan Smith, highlighted the robust nature of a strategic forecast (as opposed to a standard baseline plus forecast), concluding that if enough rigour and challenge is entered into then reliably forecasting for a post Covid marketplace (the next 12 months) should be realistically achievable.
The five-step process below details how leading marketing performance agency RedRoute International suggest brands approach strategic forecasting, and is based on how they have achieved and continue to achieve enviably accurate forecasts for organisations as varied and complex as Heathrow Airport, Photobox Group, Money Supermarket, and many others…
Almost without exception, the last 12 months has taken its toll on business. Whether it was the forced closure of retail stores, supply chains falling apart as international border controls became more onerous to navigate, consumer demand either falling off a cliff or skyrocketing through changing consumer wants and needs, the ‘corona-coaster’ analogy certainly felt apt to many… But with a gradual opening of economic opportunities, how can businesses now forecast and plan for the year ahead? The answer, unsurprisingly, starts with an assessment of where you are today.
In purely general terms, the inherent nature of the “predictive task” is now no different from what it has been in the past, the degree of rigour and challenge however, that businesses need to apply may be significantly increased.
For some businesses in stable and mature markets, forecasting may simply have been a case of ‘baseline plus or minus’ taking a few general trends and macroeconomic indicators into account. Our panel however and their many FMCG clients, have long understood the financial implications of accurately forecasting to even within a few basis points of the most likely outcome – after all balancing supply and demand can make or break commercial opportunity.
Alongside adopting a more rigorous process, our panel points out that the degree of variability will inevitably be greater in post Covid forecasting, for the simple reason that so many things have changed all at once.
Forecasting the Effects of Marketing Factors
The first thing to bear in mind is that the impact on people, distribution channels, needs, attitudes, behaviours, and marketing practices has been quite different across each sector of the economy. Digital-only businesses have in most cases seen significant increases in demand whilst, apart from some stores such as supermarkets, retail, hospitality, and leisure have suffered greatly. This means, now more than ever, each commercial situation must be considered on a case-by-case basis. There can be no ‘universal’ or ‘one-size-fits-all’ solution. Instead, we need to logically think through how things have changed, what the implications of those changes are, and how we should adapt / modify what we know from the past about how marketing works, to help create the picture of what is most likely to happen in the future.
If we have learnt anything over the past 12 months, it’s that we cannot avoid uncertainty and should expect more changes, marketplace volatility and fluctuations in consumer demand as the situation evolves, and so adopting a test and learn, or as the panel put it “Assess, Plan, Do, Verify” approach, will be critical as we build knowledge and gradually become more certain of what lies ahead, over time.
So how should businesses assess what the coming period may look like? How can we predict the most likely outcomes and impacts of our marketing investments in such capricious conditions? The five-step process, albeit somewhat simplified here, walks through a logical approach to forecasting resulting in the ability to identify and scale the key drivers, the agility to adapt as conditions change and the foresight to prepare for unlikely but nevertheless possible events, because let’s face it, no one saw this coming but it happened, nonetheless.
Step One: Start with what you know to be true. Clearly, a baseline is needed with any forecast and today that baseline may look quite different from what has been seen in the past. Traditional B2B businesses have emerged as strong DTC players for example, so taking stock of the new shape, size, and structure of your organisation, both as a whole and as a sum of its constituent parts (including supply chain and onward distribution networks) is going to be a critical step. What is the current size and shape of your business? Have the actual consumers changed? If so, what is the new mix? Have consumer needs changed? e.g. we have seen a rise in walking and cycling? Have consumer values changed? is there now a premium on being ‘covid secure’? Have the distribution channels or routes to market changed and, if so, by how much? And how might all these changes persist in the future as / when restrictions on movements and so on are relaxed?
Step Two: Identify the drivers behind every change. As marketeers we should all buy into the idea of “cause and effect”, our campaigns (cause) are designed to elicit (effect) a consumer response, usually to buy or engage with our brand. So, if we are seeing changes (effects) in the way consumers are engaging with us, we need to understand the cause (or more likely causes) or underlying drivers for those behaviour changes. City centre coffee shops for example are inevitably feeling the impact of fewer commuters passing by on their way into work, but alongside this they should also acknowledge the potential Covid fears among those still passing, or the impact of cycling or walking to work rather than taking the bus, perhaps a different route is taken, or perhaps consumers are becoming more health aware or financially prudent so are cutting back on the morning latte and danish pastry with a different motivation in mind? Most likely it will be a little bit of all of these and some other drivers, that are collectively having an impact. Its only through knowing the full range of impacts that we can truly assess the reasons behind a change. Some will be obvious, others may be less so, and because of this brand should look to update their understanding of their customer relationships over time. If consumer groups / mix has changed, then brands may also want to reconsider the segmentation models deployed. Do pre Covid attitudes remain intact or has there been a fundamental and lasting change in the consumer relationship with your brand? Brands that invest in customer insight now and as the situation evolves will be most able to adapt with their customers and stay or become first choice in category.
Step Three: Measure the impact of each driver. No single ad has a measurable impact on consumer behaviour, but a combined and sustained campaign can change the tides, and there have been numerous examples of brand campaigns that have done just that. If we think on these same lines, the impact of Covid was not a single event, but an all-consuming and sustained campaign affecting so many aspects of life for such a lengthy period, and so attributing one element to changing a business’s fortune would be overly simplistic. Each separate driver that led you to where you are today needs to be measured and evaluated for its impact to be understood. As markets begin to unfurl the blend of drivers that remain will inevitably change, for example we are as of yesterday allowed to shop in store more freely, but will consumer confidence and willingness to endlessly browse the rails of retail return at the same rate? If we understand how all the small drivers led collectively to the big changes we are seeing today, then we can plan for different combinations as we move steadily forward into a state that is more open, but changed, nonetheless.
Step Four: Learn as you go. Through stages 1 to 3 we have built the ability to create new forecasts and hypotheses based on ‘best available knowledge’, but in ‘fluid situations’ such as the reopening of economies, the relationships and nuances that led to change, will continue to ebb, and flow as the situation develops, meaning the strength of impact for each catalytic element will also continue to change. By now however, you have identified and scaled the main drivers of why consumers are choosing to (or not) engage with your brand, so continuing to monitor and assess the most crucial elements via market research, live in-market tests, etc and continuing to revisit your forecasts will provide an ongoing rigour, strengthening the projection, in the short term at least.
Step Five: Plan for the unexpected / Scenario planning. The pattern of what will happen in the long term is naturally subject to much higher degrees of uncertainty because it is not yet know how much of changes in behaviour we have seen will ‘unravel’ once normality returns. And we do not know for how long the current non-normal situation will last. So, the best approach to cope with uncertainty of this type is to use scenario planning. Many alternative situations are worked through and the results tabulated to guide decision-making. The important thing when doing these long-term evaluations is not be afraid of outcomes that look very radical compared to what had been seen in the past. The shifts that have taken place saw 15 years’ worth of growth of online commerce take place in 15 weeks in 2020. The one thing that you can say with certainty is that the future will, forever, look radically different to the past but the process outlined above will enable us to project future marketing impacts with “reasonable expectations”.
To summarise, the recommended approach to follow is one of “Assess, Plan, Do, Verify” with the five specific steps in implementation as outlined above. The inherent nature of the predictive task is no different from what it has been at any time in the past, what is different is the level of reliability that we will have in the projections which will obviously be less. We can offset this risk by putting greater effort into the Assess phase, by making sure we can measure the impacts of the things we do, and by having an effective feedback loop to Verify the outcomes and re-Plan trajectories rapidly and continuously.
For more information on strategic forecasting, analysing marketing performance, or to schedule a conversation with one of our econometric experts, please contact Sarah on ++44 7903 913 736 or visit www.RedRouteInternational.com