Payroll — You Can Laugh or You Can Cry
Background
This post was inspired by a recent discussion among the Enterprise Irregulars. It began when one of our members called our attention to a recent story about ZenPayroll’s impressive roster of successful Silly Valley (my term for the group think or herd instinct often found in the Bay Area) entrepreneurs who had become investors. His inquiry: “Hyperbole aside, this is an impressive list of Valley investors for a space that’s admittedly crowded already with a number of “next-gen” payroll tech companies already scaling as public enterprises. Wondering what everyone thinks of this company and what, if anything, is their secret sauce?”
Now, most of the EIs spend very little time thinking about payroll software and services let alone PEOs unless they’re small businessmen themselves and have to handle this as part of running their businesses. Therefore, much of the discussion came at the question posed from an investment thesis perspective, and we all know that the entire HCM product/service space is hot, hot, hot.
But I spent many of my formative professional years designing, developing, implementing and operating payroll software, and I also ran a payroll office. More recently, I’ve spent the last few decades focused on every type of enterprise HRM technology, to include payroll, so my own comments came from a distinctly different direction. And now that I’m not providing direct consulting services and can be as blunt as I choose about such matters, I thought I’d share my comments on the amount of attention that ZenPayroll (to their credit) has been generating, about the competitive landscape in which they operate, and about the relative difficulty or ease with which a ZenPayroll becomes the next ADP.
Payroll At The Lowest End Of The Market
The lowest end of the payroll market, so those software and/or services providers (and they’re all mostly a combination since there are a lot of needed services associated with payroll, from tax filing and net pay distribution to handling garnishments and benefits administration) focused on employers with 1 to 10 or 20 or even 100 employees, is highly fragmented and includes many strictly local operators running on older technology. Some of these low end providers use homegrown software but most draw upon a few quite small providers of payroll and related software designed specifically for these smaller payroll service bureaus. By and large these specialist software vendors are not in the forefront of current software capabilities (e.g. mobile, social, analytics, consumer-driven UX, embedded intelligence, etc.) nor do their Mom and Pop owners/managers present a very Silly Valley face.
A few larger providers to the lowest end of the market have PEO business models (although some of them do everything they can not to be labeled PEO because it’s so not Silly Valley) which derive most of their income from workman’s comp administration and the premiums on various types of employer-provided insurance. TriNet is the one that gets a lot of attention because it’s been built in the Bay Area, has big name investors and leaders, and gobbled up quite a few competitors to get some real scale. They also cater to the tech startup community, which tends to have very complex compensation and benefit plan requirements. Paychex has real scale and delivery capability at the lowest end of the market, and they’ve been doing a complete technology refresh to offer a more consumer-like experience while having industrial strength behind the curtains. ADP has both PEO and non-PEO offerings at scale for the lowest end of the market, and they too are bringing a real consumer-like user experience to this market. And when ADP decides to go after a market, they can afford to price as they choose, especially with their newest technology. In case you’re wondering, several other major players at the lower end of the mid-market (so starting at a few hundred), including Ceridian/DayForce, Kronos, and Ultimate Software, don’t focus on the under 100 employee market.
Thus, the market into which ZenPayroll has inserted itself with mobile, very consumer-like software, is one which has:
a few very capable larger players now refreshing completely their software (and, to some extent, their cultures and brands);
a few moderately capable mid-sized players, like Paylocity, who’ve attracted significant outside investment for the first time and at a time when if you can spell SaaS and HCM you’re a candidate for such investment;
a large number of Mom and Pop often more local operations running on either their own homegrown software or software that targets this industry from a few very small specialty software shops; and
a few other startups which, like ZenPayroll, are bringing a very fresh look and feel to the payroll experience of small business owner/operators, particularly those who already live on their smart phones.
Payroll Starts Easy But Gets Very Tough Very Quickly
Payroll is pretty easy if you’re just building sexy-looking, modest functionality, US payroll, that automates the most common use cases and uses manual intervention to handle the very occasional variation. But it’s extremely difficult to design and build effective-dated, metadata-driven, payroll engines that can handle at scale and with minimal human intervention the full range of convoluted policies/practices/regulations/etc. that companies encounter as they grow and spread geographically.
For example, if systemic effective-dating isn’t built into the foundation, then you can’t do fully automated retroactive processing. With 10 or even 50 employees, you could handle a late-breaking set of retroactive business rule changes with brute force, so with manual calculations and lots of “zapper” transactions. But with 500 or 1,000 employees, with rules late-breaking rule changes varying by states and localities and at different times, with intervening pay rate changes, and so on, it’s not possible to do brute force retroactive processing cost-effectively and, more importantly, without huge errors. And heavily to fully automated retroactive processing, where you’re able to handle intervening rule changes along with intervening software updates, is just one example of what happens in payroll when we move from the very low end to the demands of enterprise customers.
Unless you have great software that just does it right with minimal effort or you expect your customers to do the grunt work and take the error risk unto themselves, software built for the lowest end of the market may not be able to scale up — and that’s been the case forever. There are no payroll systems today that serve equally well the 1 to 10 or even 100 employee market and the 500 to 1,000 employee and above market, and the example I’ve just given is one of the reasons. Things I must automate with great expertise and cleverness to handle complexity and scale just don’t bother me at the lowest end of the market. With roughly 18,000 regulatory jurisdictions in the USA, from taxing authorities (state, several types of local, not to mention reciprocity agreements) to judicial entities which can garnish wages, payroll software for even the low end of the middle market is expected to handle all of that, all the time, and without any slip-ups.
Another important point is that mistakes in payroll processing don’t go unnoticed, and a new and/or small player, no matter how sexy their UX nor how wealthy and prominent their backers, will get nailed with their first serious error. And the opportunities for serious errors are huge. Years ago, every tax/small business accountant did payrolls for their small customers, but many haven’t been able to keep up with the technology requirements or the complexity and risks of today’s payrolls. Banks used to do payrolling on a large scale for smaller firms, but fewer of them are doing that today for the same reasons. Presumably, the ZenPayroll guys, whom I don’t know, have tons of substantive experience and knowledge and not very little hubris. And hopefully, they have on their team one of the handful of folks who know how to design and build a modern payroll engine, i.e. an engine (not an application) which is great and durable, effective-dated, metadata-drive, scalable and low cost to operate. I haven’t yet seen any coverage of this firm which has probed deeply into their software’s architecture, their technology choices, and their operational robustness, but I’d love to read any such coverage from objective and very capable sources.
Bottom Line
There’s a ton of low end payroll business in the USA and plenty of room for new entrants. Giving software away and making money from the sale of ancillary services or using a PEO model to derive income from insurance premiums are well-established business models in this market. But for those dollars to add up enough to be of serious interest to VCs looking for a future liquidity event, I would think you’d need a damn sight more than 1,000 separate customers with an average of 10 employees. Rather, you’d need the tens of thousands, even hundreds of thousands, of such customers now serviced by the larger players. And to keep those small business customers, I would think you’d need to offer a lot more than basic payrolling, so everything from background checking and other pre-hire services to other fundamental small market HR services like template employee “manuals,” regulator advisory services and some flavor of basic time keeping and performance management. Those fifty big name investors in ZenPayroll must have seen something, either in the competitive landscape or in the specific business model and assets of ZenPayroll, that has whetted their appetites. But it’s not obvious to me how this ever turns into the next multi-billion $$ acquisition or huge PE multiple IPO, which is probably why I won’t ever get the big bucks that VCs do.