Articles
What is Image Recognition Engine?
First of all, it has been asked many times about the meaning of “Retail Execution”. It is predefined series of activities at point of sale to reach one single target, a “Perfect Store”. It is a marvelous journey, long and tough as well, for that matter you will be needing your helpmate, an automatic analog to digital converter, an image recognition engine (IRE).
Image Recognition Engine (IRE) is a combination of real-time on-device Image Recognition (IR) and software ports to transfer the KPI results to a customer software (SFA, BI, ERP or even a simple database), a real-time solution alternative to the manual compilation of retail data for retailers and consumer packaged goods (CPG) companies. Data is collected as real-time snapshots at the physical outlets via mobile devices owned by merch/audit teams or fixed cameras and analytics are presented to field team and the HQ in a few seconds to take actionable decisions based on the KPI results as the field operation goes on real-time as well.
Unsurprisingly, IRE’s biggest benefit is the time saved manually compiling data. Where it was once commonplace to gather data&compile it in a spreadsheet or label the shelf images manually by operators, eventually get the KPI results approximately 24 hours later; many brands now see this as inefficient — preferring to spend this time on other, more important activities like taking action based decisions on informative real-time data visible both at the sales point and the HQ.
The second benefit of IRE is the ability to improve analytical accuracy, ultimately, your understanding of your real-time physical performance in categories across the physical world. Automated IR data is void of human error, providing you with your planogram compliance and where your products sit in comparison to competitors’.
Lastly, IRE empowers you to track the real-time physical performance and having real-time KPI results opens multiple doors such as getting snap Retail Execution decisions or taking automated orders even by crowdsourcing usage. The fluidity of this data also makes it easier to spot trends, survey price changes and make data-driven decisions form the HQ.
At Detajo, we live by Retail Execution thus the excellence of IRE.
What are the methodologies to collect KPI data by a field force at retail execution?
This is the question which each CPG company should decide for their retail execution processes. Many companies are already doing it for their products available at online shopping platforms. But it is easy, they just stream the platform data to their reporting infrastructure and create any report based on their desired KPIs because the data is already in digital form. But for the retail execution throughout the physical outlets, the first issue is to transform the analog physical world into the digital data. There are 4 ways of doing it and the companies shall choose one out of 4 different methodologies for their retail execution.And the main differences are:
1- Manual collection: Field force simply writes down the shelf metrics on a page and later at the office they are typed by operators into a database (or excel).
Swiftness: KPI results are available hours later. Impossible to take instant actions during the visit according to the real time KPI results
Accuracy: Open to human mistakes in 2 layers both at writing and typing stages, average accuracy is less than 60%
Cost: Consumes an enormous time of field force and the operators, the cost is the highest of all.
2- Manual collection using a sales force application: Field force simply fills the required fields or chooses from the multiple choices for the specific outlet at the sales force application screen.
Swiftness: KPI results are available instantly. Possible to take instant actions during the visit according to the real time KPI results.
Accuracy: Open to human mistakes, average accuracy is around 80%
Cost: Consumes a big time of field force with an average of 1 hour for each visit.
3- Automatic collection with "Image Recognition-IR" solution supported by operators in the shadows: Field force simply takes photos of the physical shelves but the IR engine is not strong enough to reach to the desired accuracy, so the provider uses operators in shadows to increase the accuracy to 95% by correcting each photo manually.
Swiftness: KPI results are available hours later because of the hidden operators of the provider. Impossible to take instant actions during the visit according to the real time KPI results, should revisit the same outlet in the future again to take an action.
Accuracy: 95%
Cost: Consumes a big time of provider's operators, thus the IR solution will be expensive for all traffic values.
4- Fully automatic data collection with "Image Recognition-IR" solution: Field force simply takes photos of the physical shelves, as the IR engine is fully AI based, the accuracy successfully reaches 95% without any operators in the shadows.
Swiftness: KPI results are available instantly. Possible to take instant actions during the visit according to the real time KPI results.
Accuracy: minimum 95% at the beginning. Increases with the time, as the engine gets more experience during the usage.
Cost: Consumes negligible time of both the field force and the IR provider. Provides huge labor cost savings for both sides thus also the IR solution will be inexpensive, with a decreasing rate according to traffic values.
At Detajo, from the time of our foundation, we had chosen the 4th methodology as our core business model and focused on its efficiency.
What is “Crowdsourcing” in Retail Execution?
Retail execution is a business process designed for consumer goods companies (CPGs) that their overall retail strategy executed in retail stores is in compliance with their brand strategies. In other terms they are designed to put the right product on the right shelf at the right time to maximize the revenue. The biggest challange is to monitor the compliance of the execution around the clock, thus the objective of a store audit is to capture the information, qualifying whether the merchandising by the retailer has been accurate against the commercial agreements agreed with the manufacturer. which is very demanding in terms of labour. Crowdsourcing is changing the ownership of this process.
Crowdsourcing related to Retail Execution is to distribute the effort required to collect data and audit stores by not dedicated Field Sales resources or third-party audit companies but instead, to appropriately place geographically local resources such as a network of shoppers, store owner engagements or even seasonal student groups.
What is “Crowdsourcing” in Sales Operations?
Sales operation of a CPG or its distributor covers 4 main areas to focus:
1- Traditional Channel such as stand alone stores or even local store chains
2- Horeca Channel such as hotels, restaurants and cafes
3- Modern Channel such as retail chains
4- Digital channel such as e-markets or online sales
Crowdsourcing related to Sales Operations covers the first two channels. Here by organizing store owners/managers to use image recognition engine platform instead of dedicated field sales force, will not only achive online order placing function by shooting the shelf and approving the automatically listed missing products thus placing the order in real time, but also to collect data and audit stores too.
Contact us
Interested in working together? Fill out some info and we will be in touch shortly. We can’t wait to hear from you!