Case 3: What does it mean when we say that the promotion is present in stores?

Adam Beniowski
Adam Beniowski
11 August 2022

Let's start with a series of very important business questions which are often asked, but not always with fully defined control KPIs, leading to incorrect conclusions and financial losses:

  • If the report says that, out of 100 contracted stores, the promotion was present in 80 of them, what does that mean?
  • If the promotion was supposed to last 2 weeks and 40 stores had it for only a week, was there a promotion or not?
  • If 50 stores had 3 products out of 4 that were supposed to be included in the promotion on display, was this promotion present in these stores or not?
  • If, as part of a promotion, a customer could get a free product at the cash register, but only 60 stores had POS materials to inform customers of that possibility, was the promotion present or not?

Everyone who has ever had anything to do with retail realizes how important it is to maintain proper exposition during promotions. We work with a number of clients from various channels, and we can assure you that this applies just as much to FMCG, Pharmaceutical, Apparel, Home Appliance, or DIY industries.

This is important both from the point of view of the retail chain and the supplier/
producer - not only because there is a great deal of investment associated with this form of promotion (e.g., according to our estimates, producers and retailers spend around 9 billion EUR a year on all kinds of promotional expositions in Poland), but also because its main goal is to increase sales. The lack of sales growth, or a lower rate of growth than expected, can significantly affect a company's financial performance.

The change in sales level, as confirmed by cash register systems, is just the end result of the promotion. What comes a step earlier in the process of assessing the execution of the promotional activity is the actual quality of promo expositions. Without taking this into consideration, it is impossible to properly determine the effectiveness of the promotion. Since it is so important to have these additional
promotions executed properly at the point of sale, naturally most aware companies measure this parameter. For most producers and retailers, it is obvious that it is worth monitoring the business performance indicators which affect the company’s financial results. Parameters which are not controlled tend to erode naturally, that is, they simply stop being implemented by stores (which these days, by definition, have more tasks than capabilities).

Whether a parameter is measured or not is one thing. Another question is what we actually get out of it as a result. Often the method of measurement itself dramatically reduces the level of accuracy. For example, our client has been measuring the percentage of correctly executed expositions for many years using conventional methods, just like many other companies. The measurement was based on external auditing companies, merchandisers' reports, sales representatives' reports, or simply believing that the retailer's operational structures are able to carry out contractual arrangements to the letter. The important thing is that all of these methods are based on a declarative form, with no possibility of secondary verification of the actual situation.

Thanks to Artificial Intelligence algorithms supporting Photo Recognition technology, we were able to work together with the client to shift to an entirely different level of exposition measurements. What is the fundamental change? Instead of human observation, it is algorithms that analyze photos of the exposition, verifying parameter by parameter. That is because the correct promotional exposition should meet several conditions at once, such as:

  1. TIME: It was visible in all stores, at the right time. Thus, it was present from the
    beginning of the promotional campaign until the final day.
  2. ASSORTMENT: It included the full assortment of promotional products previously agreed upon.
  3. COMMUNICATION: It communicated the promotional mechanism to the shopper in the way that was intended by the authors, e.g., if the price is one of the elements of the mechanism, then its maintenance and communication are also important.
  4. POSM: It included an appropriate set of POS materials, which of course support this communication. POSM can also play an additional role, such as communicating a set of brands which were part of the exposition.

Only the compilation of this information provides an answer to the question of whether the promotion was properly executed. It is worth knowing how many stores - and which ones - were missing something, as well as what was missing. We can then consider what to do differently next time. What didn’t work and how can we improve it?

Now it is time for another question - "did the change in measurement methodology show different exposition quality results?". As it turns out, it did; quite significantly in fact. The declarative measurements collected by the client showed a level of execution of the promotion at 80-90%, while Artificial Intelligence algorithms assessed the level at about 30%. What is more, with the use of AI the client could see which specific KPIs were not met, and based on these facts, could take conscious and accurate corrective actions. Every controlled exposition was documented with photos, and the level of reliability of the measurement exceeded 99%, which ended the discussion of how it was “in reality”.

Obviously, this was not good news for a company that has been operating on the Polish market for decades… In the following months, the entire sales department, both Sales Reps and KAMs had to do their homework by taking these conclusions into their discussions with the chains. The average implementation level of correct expositions rose to 95% within a mere 10 months of cooperation. Clearly, the change in approach to how they control the quality of expositions has translated into an increase in the efficiency of the client's investments in the long term.

For better insight into the effectiveness of using Photo Recognition, below you can find a summary of how the execution of promotion improved over time within different chains that our client cooperated with at the time. Their names have been concealed, but they are not important from the perspective of what I want to illustrate. As you can see, chains X, Y, Z, and V handled the challenge of adapting to the new requirements the fastest. This was partly thanks to the quick feedback - our client was able to provide them with data from our reports as soon as the second day of the promotion. Chains X, Z and V ultimately reached a level of almost 100% of properly executed expositions. Chain Y has significantly improved its operational efficiency, with the maximum level reached during this time exceeding 80%. I can reveal that in the following months it increased further, up to over 90%. Unfortunately, chain W did not improve its results quite as much, but managed to improve them by "only" 2-3 times. For the client, this is added value anyway, since it allowed them to clearly determine the level of execution efficiency of this chain, obtaining arguments for negotiating the level of investment.

Thankfully, chain W was not sufficiently large to significantly decrease the average % of correct expositions achieved by the client after 10 months of cooperation. The overall average score was high, at 95%.

And how about you? How do you gather information about how well your promotions are implemented? How much confidence do you have in conventional reports, now knowing how they are created?



TAGS: photorecognition, prodisplay, retail, ai, artificialintelligence, promotion