Why Payments Companies Should Eventually Buy All Meaningful POS Companies

Facebooktwitterlinkedin

People seem to forget just how important data is in today’s world. Google, a $550B advertising company, is a data company. Facebook, a $350B social network, is a data company. Amazon, an ecommerce juggernaut pounding retailers, is a $350B data company. Companies that aren’t data companies at their core desperately want to be – and for good reason.

Tomorrow’s currency is data. Smart companies are now repositioning business models to capture and make use of data instead of billing customers for the main offering. I’ve never given a dollar to Google or Facebook directly but I’ve undoubtedly made them thousands of dollars over the years.

The market is rewarding such thinking in both revenues and revenue multiples. Below is a graphic of the largest companies over the past 15 years. Notice a trend?

There’s a LOT of data produced in the physical world but hardly any of it is put to use. Mostly it’s because the market is focused on other things. From the merchant’s perspective, there are 1,000 fires to put out on a daily basis. Business owners set-up-shop to offer goods and services to customers, not analyze data.

Parties that serve merchants have been focused on endeavors that bear the most revenue. POS and payments companies traditionally saw more money from non-data services, so data was merely a distraction. We’ve seen some of this starting to change…

  1. Merchants are becoming more comfortable using data to better manage their operations. Tech startups are showing merchants what’s possible with improved data analysis and the serious impacts it has on top and bottom lines. In the next decade business owners will have grown up using technology backed by data and will expect much more sophistication from their providers.
  2. POS companies are offering more reporting and “above store” features than they have previously. Investing development resources into data tools, even if rudimentary, has become table stakes.
  3. Processors have started monetizing aggregate payments data. Processing revenues are falling as competing acquirers beat each other up on basis points and new sources of revenue are being looked at more closely than before.

For the markets to wake up and prioritize data revenues, we’re going to need the right ingredients merging.

First, POS will need to continue its commoditized decline. POS prices should continue falling for the next few years as they follow in the footsteps of the payment acquiring business. With more US acquirers building/buying/partnering to offer a POS solution, you’ll see processing revenues offsetting would-be POS revenues, undercutting market prices. Continued downward pressure will be exerted by lower POS software development costs and falling hardware costs.

Second, as stand-alone POS companies can no longer make much margin, the smart ones will seek other revenue streams. Since cloud POS is making item level SKU (stock keeping unit) data available from one API for all of its merchants – a serious problem for non-cloud, local-server POS systems – these companies will be the first to benefit from the data transition.

Assuming enough item level data can be cobbled together (a very non-trivial problem in the fragmented world of brick and mortar outside grocery), the cloud POS companies will start tapping into the large co-op marketing budgets we’ve discussed previously. But they won’t get to billions in new revenue over night.

The data revenue will start at the lowest level of data: syndicated data. Syndicated data delivers aggregate insights across market swaths. For example, syndicated data would tell you that product A sold more than product B last month. Further filters on syndicated data could show you things like what geographies product A performed best in, and at what prices.

As POS companies start to discover that data must be cleaned and organized to grow in value, you’ll see POS product development changes that force merchants to more uniformly code things like discounts and promotions. You might even see heavy standardization of menu items from drop down lists, so merchants don’t have the opportunity to misspell “pizza” for instance. From here, the value of syndicated data increases another notch: can you analyze discounts and promotions to determine if they’re creating lift on item sales?

This cycle will continue as more item level data comes into the syndicated data pool and cleaner data attributes are added to transactions. Even if legacy POS companies do not make the necessary updates to survive, cloud POS providers will continue to grow market share and add data. At some point, however, the syndicated data market will reach its maximal value. At that juncture, customer data must be added to continue climbing the revenue ladder; it’s data that POS companies have neglected to collect. Here’s how that happened.

In our humble opinion, PCI compliance is a masterful ruse pulled off by the card networks. To “reduce fraud”, the PCI council – which functions like a body of the five mafia families in Visa, Mastercard, Discover, Amex, and JCB – pushed through onerous requirements for cardholder data downstream. In effect, the merchants and the card issuing banks must foot the bill for any fraudulent transactions.

If I may quote a payments colleague, “Enacting PCI compliance standards is like trying to prevent bank robberies by ensuring the tellers are dressed appropriately. My assertion is bolstered by the fact that most breaches have occurred with merchants who were recently deemed PCI compliant.” PCI also has an added side effect of limiting the number of parties who access personally identifiable information (PII) coming from customer card transactions.

POS software developers responded to the PCI guidance (it’s guidance, NOT law) by changing their software to dump all customer card data, in effect removing their software from the scope of PCI. This loosely translates to merchants being out of PCI scope by selecting POS software that has made this product decision (don’t hammer on the details; I’m just sharing a directional overview).

To further this thought, EMV has introduced an additional wrinkle in the customer data chain. Instead of customer names and truncated card numbers appearing, you’ll now see tokens unless you’re fortunate enough to have built your own tokenization scheme – something that only the largest processors could afford. EMV again consolidates data into the hands of the card networks, who have built large data businesses on said data already.

This is where the payments companies will be brought into the data discussion. Since POS companies have not been, and are not currently, collecting PII, they will need to involve the processors (or the card networks) to match transaction timestamps at the merchant check level with timestamps from the processing stream, marrying the two data sets.

I can’t emphasize the importance of joining both data sets to tap into the full potential of co-op advertising budgets. Taking the data separately, here’s what you’re given:

POS Data Stream: SKU Data. Useful for basic syndicated data. Maybe tens of millions in potential revenues at scale, but no POS company will pull this off independently.

Processor/Network PII Data Stream: Customer Names, which can give you Demographic Data. Maybe a few million from low-margin retailers who want to know their customers and where else they’re shopping.

We can think of them graphically like the below.

But married together, the data taps into the high-margin supplier dollars because you’re combining item data (which suppliers care about) with demographic data (which consumer insights marketers care about). Now we’re talking about billions – possibly in the tens – of new revenue over the next decade.

This is much more a function of the amount of data available than the willingness of suppliers to pay for it. In other words, will SKU and PII data be available across enough verticals and geographies at a deep enough level to give advertisers confidence that the data is reflective of the overall market? This is why SKU data is going to have to come from a multitude of POS companies: no one company has enough to grab the big dollars.

The chart below helps us build visuals around the relative opportunity.

When processors are made aware of these opportunities, they will simply buy POS companies. After all, Heartland was able to buy POS companies with 15,000 merchants for nothing more than $20M. That’s a drop in the pocket for payments companies churning out hundreds of millions in annual profits.

Remember, 90% of commerce still happens in the physical world, but lack of sophistication has made tapping that data extremely difficult. When, not if, that changes, there will be all sorts of new Googles, Facebooks and Amazons minted. The payment companies could make sure it’s them.

Facebooktwitterlinkedin