How AI can help drive growth for FMCG brands in brick-and-mortar retail

A quick thought experiment

 

Let’s play a game. Pretend you are a brand manager at a large FMCG organisation. Your company is paying hundreds of thousands of dollars for shopper data. It’s great. You can find out all sorts of insights from the real-world behaviour that will leave your ad agency’s made up ‘personas’ for dead. 

And as a devout attendant to the church of Ehrenberg Bass, you know it’s critical for you to improve your brands mental availability and attract more light buyers. 

But you’re faced with two painful questions.
How can I use this data to achieve my marketing goals?
And more importantly, is it worth it?

A database of almost every single grocery product.

The Shping app hosts a database of almost every single product listed in Australian grocery stores. The information is constantly refreshed as the data is pulled from various sources on a regular basis.

But that’s not the fun part.

This database is supercharged by a machine learning algorithm paired with receipt-reading technology. In simple terms, the software makes use of “Optical Character Recognition” (OCR), the same technology that allows you to search words into your Apple Photos and magically find text-based content in images.

The user snaps a photo of their receipt, the software has the ability to identify individual characters, and then AI extracts meaning from the text to recognize what each part of the receipt actually is, and how that relates to a product. All in a matter of seconds.

Combining machine learning with crowdsourcing?

The output of every receipt upload is mapped back to the product database. This starts to build an interesting picture of household purchasing behaviour.

But as you can imagine, with 20 million products across several retailers, there’s bound to be some challenges. Retailers use codes and abbreviations to list products on receipts, and in some cases the AI can’t recognize the item back to a product in the database.

But instead of just giving up completely, the app has taken a page out of Google’s “reCAPTCHA” book and users review, scan the barcode of the product and correctly map it back to the database. So that next time a user scans the same code, the system knows what it is. Pretty clever, hey?

At a large enough scale and a long enough time line, the machine learning will improve exponentially and the data will map perfectly. All thanks to a little help from the humans to the machines.

So how does AI help marketers?

What about our data-swamped brand manager from earlier? What good is more data? The answer is “not a lot”. But the hero of our story will not be dismayed, as the data is extremely useful because it has real-world applications through the Shping mobile-app. 

Run a digital ad, track an in-store purchase 

Shping partners with brands like Pepsi, Heinz, Bondi Sands and more to deliver marketing campaigns off the back of this data. Shping combines real-world shopper data with engaging ad products, giving brands a powerful tool to recruit more shoppers. Knowing exactly who’s shopping your category, what they buy and how often means you can hit them with a timely message or offer. A marketer’s dream!

You can push surveys to your shoppers. You can run a sampling campaign to drive trial. But better yet, over time, you can measure the effectiveness of your messaging and offers and optimize for converting more ‘light buyers’ to your brand. 

Wanna find out more? 

At Shping, we are self-proclaimed FMCG brand nerds, we love meeting new brand owners and understanding the nitty-gritty of what makes your business work.
To find out if we can help you, contact us here or book a meeting here.