Investors love verticalization because it assumes the cost of acquiring a customer can be spread over a broader set of revenue streams from various products, making the economics very appetizing. And we won’t disagree with this: it’s why we’ve even encouraged the POS ecosystem to sell more products to each customer and drive up the net retention to a figure greater than 100%.
Our quibble is precisely how a POS does so.
Vertizalization often assumes that the products in the stack are fairly commoditized, and that the real differentiation is distribution. In brick and mortar generally, distribution is undoubtedly the most expensive part of the business, and just because you have a great product does not mean merchants will flock to it, much to the chagrin of Silicon Valley investors. But we’re not convinced the products here are as commoditized as investors want them to be, either.
What follows is perhaps the most important part of this article as it sums up the underlying motives for the actions you have seen and will see in the market.
Once we get a SMB merchant on our cloud POS we have total control of data access; you can’t hack in like you could with legacy POS systems because the data isn’t available on the local machine. We don’t think SMBs care about getting the best possible stack solution but instead prefer a lesser product from their POS provider. This is because the convenience of having one vendor and, in theory, tighter integrations to the POS is preferential.
Too many POS investors
This premise only works within two fundamental constructs.
First, that the SMB is unaware of better alternatives and will gladly accept your 80/20 stack solution (or be oblivious to the fact that your stack solution might actually more like 20/80) over alternative products. This isn’t a bad bet since merchants don’t do any self-discovery. We’re not sure demographics change this either, meaning we don’t know if the more tech-literate generations of merchants that replace today’s luddites will be any more interested in their success than their predecessors, who seem rather apathetic to the idea that they’re supposed to be running a solvent business.
Second, that the cloud POS can make superior “alternatives” sufficiently expensive that a merchant who does happen to fumble his way into self-discovery with a drunken Google query chooses the stack solution offered by the POS instead. Remember, there’s nothing preventing any of the cloud POS companies from arbitrarily deciding that whatever ___ stack solutions will now have ___ integration fees because the POS wants to saturate their merchants with their own version of said stack solution.
More tangibly, a cloud POS builds a new labor module and all third party labor solutions must now pay a $100/mo per location tax to the POS company. That third party isn’t exactly going to eat those COGS and the merchant isn’t going to pay twice or three times as much for a third party labor solution, especially when the POS sales rep is gleefully promising all the things their labor solution can’t do anyway. Don’t think it can happen? How about this for a litmus test: if a POS will rape you on processing rates, why wouldn’t they rape you on something else? Not exactly choir boys now, are they?
(By the way this theory totally falls apart in enterprise where feature sets for stack solutions are much more robust and the bar that much higher than SMB.)
So we can agree that POS companies will likely never offer an OEM’d bolt-on solution that’s as good as a third party’s alternative. Yet that won’t prevent POS companies from trying their damndest to verticalize no thanks to investors armed with spreadsheets. If you’re forced to take a non-POS solution from a POS provider, which is likely to be the least smelly turd?
A good rule of thumb is that if you can get the bolt-on from a payment processor it’s not a complicated product. That’s because processors basically refuse to invest in R&D and you couldn’t find a product bone in their bodies. As a result, processors take on products that require virtually no upkeep (or will eventually kill the ones that do). This is not all processors but it may as well be.
In order of most to least successful, we’ll lay out the bolt-ons that we think POS companies have the highest chance of pulling off.
Loyalty
Basic loyalty is well, basic. At its core loyalty is a way to spam customers who share their contact information with offers, usually of the discount variety. There’s no intelligence to this and it’s a product any decent software engineer could build over a weekend. If you’re looking for a Mailchimp campaign product, a POS company’s loyalty solution will fill that void.
But proving if the loyalty program is driving lift, offering a rules engine around promotion types and customer cohorts, and even using AI to target specific offers to specific consumers at the item-level and across channels is way beyond the capabilities of any POS company’s loyalty program that we’ve seen.
Case in point, Heartland’s email marketing tool.
If you want a plain email marketing tool you can take what the POS offers for loyalty. If you want a real loyalty program, think again.
Online Ordering
In theory this should be the easiest bolt-on for a POS to offer. Think about it: the POS already has the menu data, they have real-time pricing changes, and they should have an API endpoint. Yet companies like Olo and Chowly exist.
Handling menu taxonomy is just not that simple. Olo has been working on it for 14 years and raised $80+ million to be one of the few successful companies in the space. They understood the menu structures, integration difficulties, and where to partner vs build. POS integration is still one of the most difficult components which is why companies like Chowly exist.
When we look at the tech landscape today, however, there are enough plug-ins and open source packages that online ordering is becoming easier – it’s why we think companies like Olo are building their own loyalty platform: they can see that online ordering is becoming easier to stand-up.
Loyalty companies, marketplaces, and others all see the gap in the market and there will likely be a plethora of new online ordering solutions in the following 6-18 months. We expect the next wave of solutions will be more scalable and we’ll even see some POS systems build usable online ordering products.
So in theory, POS companies with product culture (i.e. not those owned by a payments company) could build online ordering solutions, but we’ve not yet seen a POS do it as well as a third party just yet (maybe Toast is close, but it’s taken them $500M to get there).
Reservations
We’d argue that the barrier for a reservation tool is less technical in nature. Assuming you have an interface to the POS, the real difficulty is in driving consumer demand. This is why, as we’ve said publicly, the best bet for POS companies chasing this solution category is partnering with other POS companies and using their collective merchant base to draw in consumers.
TouchBistro seems to have the right technical approach, leveraging their relationships with Google to supplant OpenTable on a merchant’s Google listing in favor of a cheaper reservation option for TouchBistro merchants, but now TouchBistro must think about doing this across many other search interfaces. Staying on top of this might prove to be difficult, and the reality is that most merchants are deathly afraid of dumping OpenTable as they view it as a revenue generator. We’re not convinced that consumers won’t adopt other platforms – like Google – over time, but it’s an uphill battle to move merchants off existing reservation platforms.
So while the technical merits might make sense for POS to take on reservations, it’s the consumer adoption that could still be lacking.
Payroll
Everybody and their brother is rushing into payroll. Why? Because it’s relatively lucrative. Still, that doesn’t mean it’s easy.
There are a lot of intricacies when paying people while a POS transaction is comparatively simple: there’s a cost for an item and applicable taxes. Payroll requires one to wade into a minutia of regulations around taxes, benefits, minimum wages, time off, working days, paystubs, banking integrations, and more. Delivering physical checks? Add another layer of complexity.
Payroll, a bit like labor scheduling, is a compliance treadmill that nightmares are made of.
Analytics
If there were ever an amorphous product category this would be it. Is analytics just reporting? Business intelligence? Something else? Nearly every POS offer some canned reports. Business intelligence tools are rarely used because you need analysts to ask meaningful questions and it’s hard to find people smart enough to ask the right questions for $70K a year in retail. Also, SMB can’t afford to hire analysts anyway.
Most people have assumed this category to be “reporting” but that’s changing substantially. With the proliferation of data science there’s a lot more to reporting than just showing average check by employee (which we’ve shown to be incredibly dangerous when not appropriately handled by data scientists).
We think POS companies can offer basic reporting but getting to more advanced techniques like prescription or even automation is a pretty big ask. Lightspeed partnered with CrankLogic for their prescriptive capabilities, then acquired the company. Vend POS tried their hand at the same idea with Dott, but even internal people weren’t exactly thrilled with how it turned out.
Labor/Scheduling
This is getting close to the realm of, “No way, Jose”. Assuming you could build a mobile application and use the latest deep learning models for forecasting, you’d still need to monitor regulatory compliance 24/7 thanks to overzealous legislators.
POS systems can alert management to overtime and enforce clock-ins, but building a full bore scheduling app feels like a really, really far stretch. Labor is an incredibly complicated product and we’ve even laid out the specifics in a previous post on verticalization.
Inventory
And now we’ve reached the realm of the virtually impossible. The most a POS should do is a simple ingredient/item entering and deductions. For foodservice that still assumes the merchant builds their recipes in the POS, which they don’t. As we ingest lots of POS data we can tell you that only about 3% of restaurant merchants use their POS for some form of inventory. When the merchant gets large enough they turn to a third party for a real inventory solution.
If a POS were to do anything in foodservice it should be a basic produce and proteins count that hacks a rough COGS. But 85% of merchants don’t know their own COGS so even if it’s built it likely wouldn’t be used any way.
Toast has taken a broad brush to inventory, likely because they realized it was a major component of cost control for restaurants, but we’ve not heard any positive user experiences. We suspect that says more about Toast’s SMB customer base than it does their product, however. Still, we’d bet there would be a lot of holes when stacking Toast’s inventory solution against that of a company like Crunchtime!
Retail is whole other ball of wax, because it’s not as simple as an addition/subtraction of inventory. For one thing, inventory is a moving target if the merchant operates multi-channel. There are also various states of inventory status, and this is before taking into consideration inventory costing (FIFO, LIFO, etc.), unit of measure conversion, multi-currency and multi-location. Now multiply that by 100,000+ SKUs – with images – and you’re dealing with non-trivial speed and reporting issues, particularly in the cloud. Plus you have to design an import mechanism for this volume of data.
——–
We think you can take bolt-on solutions from a POS you just need to throw away the pretenses that it’s as good as stand-alone third party products. Yes, there may be some benefits of tighter POS integration with an OEM’s solution, but there’s little technical reason a third party product can’t be integrated just as tightly to today’s POS systems unless the POS system doesn’t want it to be, or the third party is incompetent.
The wrinkle to story would be if a POS company offers a white labeled third party solution as their own, or if they’ve just acquired a stack solution to fold into their offerings. In the case of the former, that’s probably the best option you can hope for. In the case of the latter, it’s good but it most likely comes with an expiration date since founders often leave post-acquisition, and with them goes the culture and entrepreneurial ethos that made the company worth buying in the first place.
To sum up, bolt-ons from POS companies are almost certainly worse than third party analogues, but they’re not necessarily bad depending what you need. But you must go into the experience with eyes wide open.
Add comment