Use digital data to maximise offline retail success.
45% higher new customer identification B2B
65% fewer delivery trips to outlets
65% fewer delivery trips to outlets
Significant reduction in out-of-stock risk
More efficient management
of your sales force
Brick-and-mortar stores are far from being as dead as is often predicted. With 2.7 new store openings per any stores closed, the balance for analogue trade, for example in the USA, it's not bad at all despite the strong growth in online competition. Nevertheless, particularly in relation to the use and analysis of relevant data, stationary retail lags well behind online competitors. While e-commerce providers use detailed data from web analysis, CRM and Google searches to analyse the behaviour of their users and develop personalised offers, utilisation of such data for a large percentage of classic retailers and brands is still a long way off.
If retailers or brands in the analogue world want to know which location is best suited for opening a new outlet, how to optimise their product range or how to use their limited sales resources in a targeted manner, many of them still rely solely on static structural data or the implicit knowledge of their employees. In spite of the extensive availability of digital geodata, strategic alignments in offline trading are often still carried out blindly.
TD Reply has developed a scalable approach that predicts sales for various points of sale, segments outlets based on their sales potential and target group fit, and optimises marketing measures per store so that stationary retailers can also jump on the bandwagon of data-driven decisions – all on the basis of a wide range of digital data. This approach which TD Reply calls "Advanced Outlet Analytics", can be applied to a wide variety of industries: food and beverages, fashion and automotive are just a few examples.
As a sales representative in the B2B sector, you are faced with the challenge of targeting exactly those outlets where the product on offer will sell best. The problem is similar for B2C retailers: which categories and products have particularly high potential per store, and how much stock should be held? The data that can be used to answer this question is initially proprietary data (owned data) – for example, first-hand sell-in or sell-out data. For brands that do not have sovereignty over this data or only have insufficient data quality, the purchase of sales data via digital POS providers is also conceivable (so-called paid data).
However, even if data on product sales is available – be it via the utilisation of existing POS data or an external purchase of such data – only a few companies have an answer to the question of which external factors influence sales success, and which are the most effective levers for maximising success. Answers to these questions can be found – often unnoticed by offline retailers – on the internet.
Services such as Google Maps, Facebook or Flickr provide large amounts of geo-specific data with high informational value. Using this, the stationary trade can put itself in the shoes of the end consumer in order to identify the most relevant information for its industry. Consumers, for example, can find out online about opening hours (Google), user ratings (Google, TripAdvisor, Foursquare), shop-specific offers (Yelp), directions (Google Maps) or relevant events (Facebook). This behaviour generates geo-specific data that can be used individually for each outlet and scaled for the entire market. It is therefore possible to obtain detailed data about the area around an individual outlet – for example about the density of shops, relevant places such as concert halls or schools, or proximity to the underground or suburban railway stops – solely via the data from geo-services such as Google Places or Open Street Map. The integration of this data together with the company's own figures represents a decisive and thus far missing link in order to make detailed assessments of the relevant consumer target group, the respective sales potential and the optimal marketing measures for each outlet.
One application case for the intelligent use of online data is the better forecasting of sales, which can achieve an accuracy of over 90 percent in the food and beverage trade. In addition to sell-out data for each outlet, the data base includes information from Facebook on local events, Google data on regional search interests and live weather data. Based on these more accurate data-based forecasts, the number of delivery runs can be reduced by up to 60 percent while minimising the risk of out-of-stock situations. Visits by sales representatives can also be better planned, which makes a decisive contribution towards increasing cost efficiency.
Another approach to using external online data is data-driven outlet segmentation, which is relevant to both retailers and brands in terms of assortment and marketing optimisation. On the basis of information about restaurants, bars and hotels in the area (e.g. via Trip Advisor), the frequency potential of the location (e.g. via information from Google Places) or the attractiveness of an environment for tourists (e.g. via the number of photos uploaded to social media), predictions can be made about turnover potential and target group fit in order to prioritise outlets and assign them to individual brands and products. If this segmentation is combined with information on marketing measures and sales data, it is possible to prioritise where it is worth taking action – be it through promotions in shops, price campaigns or a change in the product range.
Utilisation of the latest technologies, such as machine learning and artificial intelligence, plays a central role in making these approaches scalable by optimising them on a continuous basis. In doing so, it is essential that the developed models are dynamically updated and that they react to changes in data as well as environmental factors.
Nevertheless, new technologies are no cure-all for entrepreneurial tasks. New data and technologies can only be used meaningfully if they are linked with specific questions and the aforementioned implicit industry knowledge. What's more, great importance should be attached to the iterative verification of the generated results in the field and the subsequent optimisation of the models created.
"Advanced Outlet Analytics" is not merely a statistical approach that has been developed in a quiet chamber as it requires a continuous comparison with reality in the market and the existing know-how of local sales and store managers. "Data science" and "machine learning" are the buzzwords of the moment, but both can only unleash their full effect when combined with an entrepreneurial foundation.