Where Retail meets Tech

October 18, 2018

A Picture is Worth 1,000 Store Checks | Image Recognition for CPG Firms

Written by Cécile Garrett

Do you still doubt the relevance of Image Recognition for Consumer-Packaged Goods (CPG) companies? We know that it is an intriguing yet mysterious technology for still 69% of CPG firms as they question its capacities to help them improve execution and optimize return on their Perfect Store strategy (POI, 2018). But, they would be wrong not to challenge this perspective. As numerous studies and our clients attest, Image Recognition technologies can change the face of the retail execution game.

Audio: A Picture is Worth 1,000 Store Checks


Mind the Gap Between Perfect Store and Reality

The Perfect Store is at the core of all strategic retail decisions. It is designed to provide shoppers with experiences that exceed their expectations to reach sales targets. Thus, the Perfect Store represents the ideal store set-up every Consumer Goods business strives to implement and maintain to drive retail excellence and in-store performance.

According to a POPAI study, 71% of grocery purchasing decisions occur at the store. That means that out of 10 shoppers, over 7 do not have a clear brand preference when they arrive at the point of purchase. Not only that, but research has shown that 95% of buying decisions are subconsciously driven by emotions, which adds on to the pressure CG firms face to generate sales growth at the shelf (Inc., 2017). Achieving sales traction by leveraging human impulses through in-depth shopper understanding and consumer-centric marketing activities is crucial. That is why Consumer Packaged Goods companies invest substantially in marketing activities each year. Indeed, the CPG industry is the only one to allocate as much as 24% of their yearly budget on marketing activities (The Wall Street Journal, 2017). Of this percentage, CG firms spend nearly a quarter on shopper marketing activities aiming to appeal to the consumer at the point of sale (Cadent Consulting Group, 2017). 

With that much investment at stake, CG organizations must nail retail execution and Perfect Store compliance down to a T.  Otherwise, they lose competitive advantage and do not optimize their ROI. In a previous blog post about how CPGs can reduce lost sales with an Image Recognition system, we explained that three retail execution-related gaps could negatively impact CPG firms’ store sales. One of these gaps deals with the inefficient shelf monitoring and optimization processes linked to manual store checks. It affects the credibility of CPG sales forces towards retail stores and can ultimately hurt the relationship they have with them. According to POI, 73% of FMCG companies state that retailers don’t give them sufficient access to their stores. However, the report notes that “as sales reps bring insights to the retailer and are able to make decisions that impact that store, they are more than welcome.” In other words, for 73% of CPG firms, the more their sales reps can provide stores with reliable insights at the shelf level, the more trust they will generate and the greater access they will get to other POS.

Hence, generating reliable data can help CPG companies:

  • Fix planogram compliance issues to drive sales growth
  • Improve their relationship with stores by working with them towards reaching retail excellence


PlanoCheck : Collect shelf data in real time and realize the Perfect Store


It All Starts with Accuracy

Getting instant and accurate data to provide stores with meaningful insights which they can leverage to maximize revenue is a strategic aspect for CPG businesses. Why? Because the success of this task depends on the veracity of the data they share with their clients. Precisely, in an article about the power of deep learning in retail execution, our Research Director, Mickaël Maillard, explains the unprecedented degree of accuracy that CG firms can gain from an Image Recognition system:

“Deep learning systems such as the one we use in Image Recognition for retail can be fed with many good examples of the targets, under all sorts of conditions (…). [They] learn to self-adjust and recognize new examples (...) they have never seen. Without the need for explicit rules, they can recognize their targets even with low-quality inputs. They are generic, fast to roll out for new product bases, adaptive, and robust to real conditions to get truly accurate results.”

In a discussion paper about AI for businesses, McKinsey states that AI offers FMCG companies great retail execution opportunities, particularly on promotions, assortments, and replenishment. It can help them make unbiased data-driven decisions by eliminating many levels of manual activities. Vinit Doshi, a Senior Expert from Periscope by McKinsey, also comments on the power of AI in an article published in CPGmatters:

“AI doesn’t grow tired or makes mistakes; it will dispassionately crunch data with brute force (…) [and is] particularly adept at searching beyond the boundaries of our biases. (…) No longer do the restriction of human fallibility and time need to influence the process.”

Higher accuracy empowers CPG companies’ sales reps to understand and assess the reality of the points of sale more efficiently to improve sales and maximize revenue. At Planorama, we know that accuracy is at the root of advanced retail execution. Every month, we provide them with reports which demonstrate Image Recognition’s ability to deliver accurate results consistently. During pilot phases, we can also help them benchmark manual against digital measurements and assess the power of Image Recognition. This is what we did for a sample of clients between 2010 and 2017, and the results exemplify how much value CG businesses can add to their retail execution processes with an Image Recognition technology. We record that manual store checks have an accuracy level that fluctuates between 60% and 83%, versus 98% with our IR solution, PlanoCheck. That means there is a difference of 15 to 38 points between manual and digital measurements.

Such an accuracy gap reveals inefficient POS operations due to manual store checks. It is a serious issue because inefficient store operations cause the highest lost sales for CPG firms, according to an MIT Supply Chain Management research. Given the level of investment it takes to create and implement a Perfect Store initiative, losing this much money on faulty manual measurements is jaw-dropping! Do not get us wrong. We are not blaming sales reps. In truth, retail execution gaps linked to manual errors are often beyond their control. Such errors are mainly due to human cognitive biases, which can only be defeated if we substitute the heavy lifting with technology. Our Project Director, Nicolas Bosshardt, explains:

“Most of the time, manual errors come from what we cannot see: for example, if sales reps only focus on star SKUs to gain time, they might not detect if other SKUs are on the shelf or not. If products fill the entire shelf, it might look like it is planogram-compliant while it isn’t. They do not realize that there are voids.”


Close Perfect Store Compliance Gaps and Increase Sales with Image Recognition


Relying on unbiased data from store checks to close compliance gaps more efficiently is crucial for CPG firms to reach retail excellence. It is still a severe struggle. The POI 2018 TPx and Retail Execution Report notes that 80% of FMCG companies suffer from data quality issues and 90% experience compliance issues.

80% of FMCG companies suffer from data quality issues and 90% experience compliance issues according to POI

The positive impact Image Recognition has on these metrics can influence CG firms’ sales. We addressed that in an article about the toll some retail execution gaps can have on lost sales, but do you remember how big a difference it can make? Let us jog your memories.

According to Accenture, inefficient retail execution processes eat up 14% of CPG businesses’ total sales. Of those, nearly 65% are caused by Out-Of-Stocks (OOS) (Nielsen, 2018). Since planogram compliance decays at a rate of up to 10% a week according to Nielsen, CPG companies are under pressure to fill compliance gaps as quickly as possible. If they succeed in closing all compliance gaps, they can increase their total sales by nearly 8% (Kantar Retail, 2015). For instance, “we know that a 1 percentage point increase in Share Of Shelf positively affects sales by 3 points and Image Recognition solutions are powerful levers to support this dynamic,” explains Christophe Blot, our COO.

The deep learning algorithms that fuel the best Image Recognition solutions on the market can help CPG organizations optimize retail execution processes. In a paper promoted by SAP, IDC’s Vice President, Ivan Ortis, notes that “AI can support decision making with more accuracy, confidence, speed, and agility (…) without bias.This, precisely, is what makes a picture valuable for CPG firms. At Planorama, the way our Image Recognition solution can pinpoint barely noticeable compliance gaps is one of the first benefits our clients observe when they use our solution.

“When a country is moving from manual to digital measurement, we first observe that On Shelf Availability falls drastically, very often from 5 to 10 percentage points. This is because delegating planogram compliance checks to a technology which makes close to zero mistakes results in regaining true visibility over the shelf. In reality, OSA is lower than what manual compliance reports state. For the first time, Image Recognition provides CG firms with an unbiased picture of the situation in stores as well as with the necessary insights to act on compliance issues instantly,” says Nicolas.

In the following weeks after implementing digital measurements into retail execution practices, we note an increase in OSA. It is due to the combination of accuracy and speed which helps our clients’ sales representatives make the most out of their visit.

Would you like to know more? Talk to our sales teams now and start experiencing the power of Image Recognition for CPGs today:

PlanoCheck : Collect self data in real time and realize the Perfect Store

Topics: image recognition, retail execution