Keeping retail performance steady calls for sharper measurement
Retailers have stopped being blindsided by sudden shopping shifts and now recognise that today’s strategies and campaigns can no longer be based on yesterday’s spending behaviour
During the early outbreak, consumer focus switched firmly towards ecommerce, with digital buys soaring by over 20% to hit a 10-year high. In recent months, however, the pendulum has swung again — bringing the highest in-store sales since 2016 and significant online declines. For retailers trying to maintain strong relationships and revenue, this makes plotting their next steps highly challenging.
From here on, it’s clear greater agility will be essential. To keep up with fast-moving shopper habits, retailers need sharper measurement that not only allows them to track and optimise real-time performance, but also find which approach will fuel the best results in any situation.
Proxies aren’t enough to track shopping extremes
Although shopping trends have always fluctuated, pandemic disruption has drastically altered the rate and intensity of change — and not just in terms of ecommerce uptake. LoopMe’s own research illustrates UK consumers have gone through significant extremes of both sentiment and shopping inclinations, with some moves happening over just a few months.
Between March and May 2020, the percentage of Brits with a negative outlook on COVID-19 shifted from 67% to just 31%. At the same time, initial reduced purchasing for more than half (59%) quickly transitioned into stable – and even higher – spending, as 28% upped their monthly buys. According to the latest data from September 2021, shoppers are looking forward to bigger splurges ahead, with 47% now confident they’ll be spending more with retail brands in the next six months.
While not unexpected given the uncertainty consumers have experienced, and the recent lifting of lockdown rules, these findings demonstrate that retailers can no longer afford to be held back by outdated analysis. Continued reliance on proxy metrics, such as click through rates (CTRs) and impressions, means many are still grappling to make decisions using imprecise understanding of return on investment (ROI), as well as lacking the full performance picture needed to enable efficient adjustment.
Aligning with new trends as they unfold requires an ongoing supply of accurate insight. Tracking campaigns against outcome-focused KPIs from the beginning will allow retailers to assess the impact of every ad on core goals – from boosting brand affinity and purchase intent, to in-store sales. Crucially, the data this generates about audience likes and dislikes will also drive strategic, tactical, and creative adaptation that improves effectiveness.
Maximising impact calls for smart assistance
Implementing more tangible metrics is a good start to pave the way for swift responses to evolving consumer behaviour. Increasing real-time ad relevance and resonance, however, takes further measurement refinement.
To bolster immediate results, retailers must be able to review the probable success of each ad before delivery, including assessing each contextual factor in the mix and the likelihood of whether ads will hit their mark with shoppers. This necessitates the capacity to weigh up a wide array of variables and drill through vast data stores at speed — which in turn, calls for additional support from smart machines.
Artificial intelligence (AI) has already made its way into multiple aspects of the retail world, with 40% of global retailers considering their AI usage to be ‘mature’. But with a current heavy focus on labour-saving benefits, the potential it brings to spot and target prime opportunities for impactful advertising still remains relatively under-appreciated. Joining the wave of early adopters for subsets such as reinforced learning (RL) can therefore give retailers an edge over rivals, as well as improving their ability to bolster in-flight performance and fuel goal outcomes.
RL has particular value as an optimal route finder, with algorithms trained to determine the best direction of travel in differing circumstances. For instance, one obvious retail marketing use case is streamlined ad serving. Tapping data about current and past ad requests, RL can evaluate whether ads will spark their desired actions — including store visits and online purchases — by analysing the influence of multiple factors on individuals. Equipped with an instant view of advertising prospects, retailers can then narrow media spend to ads with the highest odds of powering engagement and healthy returns.
Knowledge is the key to keeping consumers loyal
Despite recent dips, estimates of over £120 billion in spending this year prove digital retains a significant share of retail sales. From the consumer perspective, our recent findings also signal that its popularity is far from fading – 30% spend more online than in physical stores, while over 40% would prefer to make their clothing and electronics purchases via the web.
As retailers ramp up competition for online attention, providing seamless experiences is becoming an increasingly critical differentiator — especially at a time when 24% of consumers name convenience as the most important element of web shopping. But fortunately, this is another area where the right blend of outcome-centric measurement and AI analysis can help retailers fine-tune their advertising approach.
For instance, by combining incremental performance data with supervised learning, retailers can predict what is likely to work for individuals in the near and longer-term future. At a fundamental level, that might involve harnessing past data about ads that have reached or missed their target to build reliable predictive models. Going a step further, running analysis within Bayesian systems allows AI engines to further expand their forecasting capacity.
By learning how to create models — for every data set — that assess many potential influences on user decisions and evolve in line with fresh insight, sophisticated AI mechanisms can make almost any outcome predictable. Drawing on data from first-party interactions, purchases and audience intelligence research, retailers can develop an array of models, each capable of unearthing deep patterns, considering the impact of all variables, and identifying how specific audience groups are likely to behave. Using the intelligence this generates, retailers can ensure ads are persistently matched to shopper needs, interests and ever-changing contexts.
While the exact shape of the post-pandemic world remains uncertain, retailers can be sure of one thing: consumers won’t be going back to standard shopping behaviours. To keep marketing performance steady, retailers will therefore need a precise understanding of the impact ads are making and which placements present the greatest opportunities. Gaining this granular understanding will mean innovating to stay in sync with variable consumer habits – including upgrading metrics to include measures that tell the full performance picture and embracing the multi-faceted advantages of AI.