CPG companies have found the future of customer marketing in a decades-old technology
When was the last time you heard of a decades-old technology that has become a game-changer for a global industry? Would you be surprised if the next generation of CPG customer marketing could be enabled by a ubiquitous (computer scanners with OCR software have been around since the mid-nineties) technology that was originally designed for other purposes?
Get ready, because as you read this quick history of OCR technology, you are likely to experience a head-slapping moment sparking the thought “why didn't I think of that?”
Optical Character Recognition (OCR) technology was introduced in the early 1900’s, and from its inception to the present day it has revolutionized the way we process information. The technology, which enables computers to read and convert printed or handwritten text into machine-encoded text, has revolutionized data processing and information retrieval. Today, OCR is an essential tool in various industries, from banking to healthcare.
The groundwork for OCR was laid by Emanuel Goldberg, a Russian-born inventor who developed a machine capable of reading characters and converting them into telegraph code. Although primitive by today's standards, his invention marked the beginning of OCR technology. Reading Machines for the Blind were introduced in the 1930’s as researchers sought to improve the quality of life for the visually impaired through devices that could scan text and convert it into audible speech.
Much later, to improve the accuracy of OCR systems, the American National Standards Institute (ANSI) introduced two standardized fonts, OCR-A and OCR-B. These fonts were designed to be easily recognizable by both humans and machines, streamlining OCR performance. This paved the way to the introduction of the Kurzweil Reading Machine in 1976.
American inventor Ray Kurzweil is widely cited as the modern “inventor” of OCR, but to be fair, Kurzweil is the person associated with the commercialization of OCR through his business dealings with Xerox and the flatbed scanner. In 2005, HP Labs marked the beginning of the integration of artificial intelligence (AI) with OCR technology by development of its Tesseract OCR engine.
But the tipping point for consumer recognition of OCR is linked with the launch of the first iPhone, introduced by Apple on June 29, 2007. The widespread adoption of smartphones and tablets made OCR technology more accessible than ever as mobile apps allowed users to scan and recognize text on-the-go, transforming the way we interact with the written word.
Running in parallel with the evolution of OCR as a useful business tool, was the continuance of the longest standing challenge in customer marketing for Consumer Product Goods companies (CPG). The problem surrounds customer identification. Working in a distributed sales chain, CPG’s Proctor and Gamble, Unilever and every other product brand on the planet were able to track sales by product type and by sales outlet (for example, a grocery or convenience store) but they could not connect their sales data with an individual customer who purchased their products. Customer identification was impossible for CPGs until recently.
The big story here is not just that methods exist to connect individual customers with product purchases. Now, we are at the point where the technology is highly accurate, cost-efficient, and prepared to bring added productivity to the companies which need it the most.
The current wave of Digital transformation has all marketers thinking of how to connect people, processes, and systems elegantly to drive better business results. Most businesses are seeking to deliver better experiences for their customers and realize they can do this by developing stronger insights into the data they collect, understand customer preferences and in turn, drive more delightful and valuable experiences to customers. Personalized customer experiences are expected by customers today and developing brand loyalty is contingent on fine-tuned campaigns that can only be generated through real-time, data-driven customer insights.
For decades CPGs have been looking for ways to connect the identity of individual customers with the products they buy. Do you remember the codes on cereal box tops? Well, CPGs quickly recognized that a better model was to at least force the customer to purchase the cereal to find a code or token inside the box. Customers followed the cue, but the process of accumulating tokens and submitting them to the CPG through the mail was clumsy and time consuming. The modern consumer is time-starved and old models had to be redrawn.
Enter receipt scanning with OCR technology. Customers are now able to scan a purchase receipt and submit the result via their mobile app. Loyalty program members can scan receipts on their phones while on the move, saving time and earning rewards. This is both a time saver for the customer and a big win for CPGs. The fact that consumers no longer need to be in front of their desktop computer utilizing a flatbed scanner to convert a piece of paper into an image was a big step forward. Not only were CPGs able to associate an individual buyer with their product, but they also unlocked a huge set of data for analysis. For the first time, CPGs are able to build visibility of customer shopping habits across their entire range of activity. But the benefits go beyond meeting this basic need.
Veryfi is one company that is pressing the forward edge of this solution for CPGs. We connected with Dave Van Everen, VP Marketing at Veryfi to learn about their approach to solving the customer identification problem in CPG marketing. Dave shared the story of how the founders of Veryfi landed on their current business model. Both founders had familiarity in CPG marketing, customer loyalty, and receipt management from their time at Coupons.com. Through that experience, they recognized the need to bring greater efficiency to receipt processing and to find ways to reduce or eliminate costly manual data entry. As it happens, both founders had Moms that were bookkeepers. Their frustration with manual data entry led them to seek out ways to lessen the burden on resources and to improve productivity. Solving that basic problem led to a bigger idea at Veryfi.
Dave commented “that while OCR is useful, it was extremely limited. The technology could read characters correctly but it wasn’t able to interpret characters into meaning. We set out to create a machine learning model that would map out a knowledge graph to close this gap.”
Building a solid machine-learning model meant that it had to work at scale. It also required ingesting huge volumes of data to train the model. As entrepreneurs do, Veryfi found a way. They launched a free expense management app targeted at consumers and small businesses. Most expenses are processed in the form of receipts and running an app that was being fed receipt data gave them the training data they needed. The approach was anything but nefarious. App users loved the functionality so much that they continually asked why the app was free. Eventually Veryfi began to charge for the app and today Veryfi continues to have a small portion of its business based on that product.
In the process, Veryfi developed an AI model that fuels a powerful marketing platform for CPG customer marketing. The engine had processed tens of millions of receipts before launching its current platform and that number has grown into the hundreds of millions of receipts today with data coming from around the world. Training the engine becomes easier and more efficient over time as the model builds proficiency in pattern recognition.
Van Everen shared the wisdom behind this approach, saying “building a business around an app has limited horizons unless you get really lucky, and it goes viral. But if you build a business around a platform, that can turn into a Silicon Valley unicorn.” That seems to be where the company is headed today.
Using receipt scanning to validate purchases in real-time is the starting point. The AI engine can interpret receipt information so accurately that fraud prevention is enhanced through detection of duplicate and doctored receipts. Costs of manual entry are vastly reduced.
But wise marketers recognize the opportunity to develop insights from the data collected to build greater trust with customers by creating more personalized campaigns and offers. For loyalty program members, modern receipt scanning can build understanding of customer preferences and generate rewards and offers in real time.
The end result is that CPGs can learn about their customers while processing information at a lower cost and improve their overall marketing efficiency. As brand marketers invest in planograms and pay for in-store product placements, receipt scanning can inform the performance of an end cap placement or a test the results of placing a product directly adjacent to a competitive product.
As OCR-driven receipt scanning platforms improve, CPGs will be able to acquire a greater share of customer wallets and serve loyalty program members in a more personalized manner. Who knew that what is considered to be “legacy” technology could be the power behind tomorrow’s CPG customer marketing frontier?