9 Jul 2018

This is why your retail data is inaccurate

DATA can provide you with valuable insights on your customers. For retailers, however, most data is out of date or totally irrelevant because 99 percent of retail transactions are offline.

This poses a huge problem for fast moving consumer goods (FMCG) companies. FMCGs need to anticipate the changes in demand at the stock keeping unit (SKU) level, as they need to anticipate demand in a timely manner.

Unlike online transactions where every transaction is tracked and updated in real-time, in-store purchases are recorded on physical receipts, while transaction data are recorded by point-of-sales (POS) system providers.

Currently, the only way to estimate this offline transactions volume is to rely on information compiled by incumbent data agencies. These data can sometimes be up to six months old, which is useless in a fast-moving retail environment.

Reynazran Royono, Founder and CEO of Snapcart said in the Wild Digital conference on Wednesday, “offline channels of shopping will never disappear. It is forever relevant.”

He explained that some categories of goods, such as cosmetics, are easily converted to online purchases. These kinds of goods are pricier, bought less frequently, and in smaller quantities.

For goods like groceries – which are cheaper, bought more frequently, and in bigger bulks – it is incredibly difficult to totally adopt an online approach.

Compounding on the problem, many transactions for FMCG products aren’t even recorded on receipts. Royono said, “More than 50 percent of offline transactions in Asia are made in traditional shops, or what we call mom and pop shops. They don’t have POS system, which means there are no receipts.”

This represents a huge portion of unrecorded transactions. From a distributor’s point of view, these are market blind spots. They are unable to provide accurate, up-to-date data to the suppliers regarding stock demands.

However, it is key for companies to be able to capture and understand these data. Estimation and past experience can only provide you with a rough margin; it won’t help you cope with any sudden market fluctuations.

Using technology like AI-powered optical character recognition, companies can translate data recorded on physical paper into digital data. The results, paired with machine learning powered SKU identification, will allow companies to fill stocks more efficiently.

That’s what Snapcart is doing. It pays end-users a small fee to take pictures of their receipts, which the company translates each item to data such as SKU names, prices, volumes, promotions, discounts, payment methods, branch location, store name, time of purchase, and total value.

Additionally, shoppers are asked to fill in demographic surveys in exchange for more rewards. This provides information such as basket content, social media, device, geo-location, etc. to better profile the shoppers.

These data are then fed to companies in real-time. What this provides, is a near real-time database, that will show how demand is changing as well as the target populations in a particular area. This will allow brands to better target their advertising as well.

“SKUs are always rotating. Every week there will be new products or promotions, as well as discontinued products,” Royono explained. The constant rotation of goods means companies need to determine stock provisions and anticipate demands quickly with the help of insights.

Of course, this relies on shoppers’ consent to provide those kinds of information in the first place. According to Snapcart’s latest download numbers, it has past 1 million downloads on Google Play Store. While not every shopper would be happy to give up data on their shopping habits, a significant amount of users across the region seem comfortable to provide that information.

Imagine if you can capture even just 10 percent of the offline shopping transaction data. That becomes a determining factor for companies to own the fast-moving retail markets.

________________________________________________________________________________

Originally written by Satoko Omata published in Techwire.asia on 9 July 2018.