Momatos’ video analytics counter to competition from e-commerce

In the retail competitive landscape, the granular data availability with e-commerce companies proves a fertile ground for them to mine customised offerings and drive sales. Not the case with brick and mortar stores. Among other factors, they have to go by what the operating staff, especially the store manager’s feedback on the reasons behind the store performance. 

With the deep pockets that e-commerce and the quick commerce companies have in India, they are proving to be a big challenge for physical stores. So can brick-and-mortar stores equip themselves with the kind of data they have about the customers, like how the ecommerce companies are able to collate?

Momatos, the kids wear retail chain operating 40 stores pan-India has attempted to do just that. The retail chain relied heavily on the information provided by the store manager and staff on the performance of the retail outlets. For retail chain stores, footfall is the single biggest revenue generation factor. In the absence of a mechanism to gauge walk-ins at stores, the Ahmedabad based retail store chain Momatos relied on the claims from the store manager. 

What typically happens in retail—in the absence of enough footfall, the store team or manager often gets blamed for poor performance. They hide behind the false assumption of low footfall or walk-ins, adjust numbers in billing to impress the head office. Momatos was looking for a solution that could provide data backed intelligence on the number of walk-ins in the stores. The company implemented Jarvis, a video analytics software that operates via the store’s CCTV cameras.

High reliability on manual claims
“We didn’t evaluate many alternatives, moving quickly to pilot the software. We started with 5–10 stores. It seemed successful and not too costly. Considering the budget, I felt this AI solution would reinforce operations,” says Nikhil Menaria, Director & CFO, Momatos Retail.

The technology gave results soon after it was rolled out in the select stores, “Once we switched to Jarvis, we saw large data variations between what the manager claimed and what the system captured,” says Menaria. 

After deployment, specific stores having good footfall but low conversion were identified from the findings. “In one case, in our store in Hyderabad (which has the highest footfall of all our stores pan-India), footfall existed but without commensurate business conversion. We started following up with the store manager about the high footfall numbers and noticed a 10–20% business growth,” says Menaria. However before the implementation, the store manager would claim low footfall for poor sales performance. “When we speak with them with hard facts and numbers, efforts follow. The store manager, otherwise, would take it as a case of pressure tactic to increase the sales number,” Menaria says.

Improving store-wise performance
The primary benefit realised after the pilot, was that the company was able to focus more on how to tackle the challenges coming in the way of low sales numbers, store by store. So, marketing efforts were increased to tackle the challenge of low walk-ins; sales staff training was enhanced, if the walk-ins were high but the conversion rate was low, etc. 

“With good results from the pilot project, it gave us confidence to implement it in all of our 40 stores pan-India. The solution is Live in 35 stores and the implementation is ongoing in 5 other stores,” says Menaria.

So is it just about tracking walk-ins? Initially Jarvis was used to track footfall. Next, “We set up tracking for customer disengagement—whether the store staff greeted customers as per brand SOPs. Before Jarvis, we had no data on greet compliance. Once implemented, I could see daily engagement percentages, which opened new avenues for improvement,” noted Menaria.

Tech behind the scenes
Jarvis captures human entries via AI integrated with existing CCTV cameras with no extra hardware cost. It identifies the number of people entering and shows a red flag, when sales staff fail to engage, in real time. When someone walks in and isn’t greeted, Jarvis sends real-time alerts based on camera data.

On a daily basis, four reports are generated, the access of which is made available to the concerned CXOs. 

1. Footfall (with gender breakdown)

2. Customer disengagement (count)

3. Customer journey (average time spent)

4. Store timings (opening/closing hours)

These reports go to store heads, zonal heads, COO, CEO, CFO—filtered by hierarchy. Reports are on-demand and live. Alerts are pushed in real time to mobile or desktop.

Staqu Technologies, the company behind the video analytics software Jarvis holds two patents for the software, “We have two patents and 25 research papers published, so far in the AI domain. The first patent is for the re-identification technology. It means the technology designed by us is able to identify a person with different kinds of camera angles without using the face. Thus there is no privacy breach of the person,” says Atul Rai, Co-founder & CEO, Staqu Technologies. The second patent is in the area of data analytics, big data search.

Staqu Technologies participated in the deliberations with the government for drafting the Digital Personal Data Protection Act (DPDP) Act. 

“We are already General Data Protection Regulation (GDPR) compliant. The law prevents anyone from identifying anyone in a way that you can get the person’s identity. What we do is re-identify you, using your apparel, your gait or your walking area,” says Rai.

“The second question is on data confidentiality or security. In that case, the data is not coming to the Staqu server. It remains on the customer's cloud,” he adds.