I took the opportunity to catch up and I asked him what's the key piece of advice he would offer an enterprise on its HR technology strategy. He answered me with a botanical analogy, explaining that an organization has to work out whether it's a sunflower or a daisy:
If you take the case of a sunflower, what this means is you have a strong focus on one central vendor. It might be SAP, Oracle, Workday or whatever, but you have one vendor you turn to for your default solution. And you surround that large entity, where you use a lot of functionality, with small vendors ... but you're relying for 80% of your capability around that central vendor.
The daisy, on the other hand, is where you have a central vendor and you essentially use that for administrative record-keeping, and you rely more on the edge vendors, or niche vendors, for the majority of your innovation and the majority of your functionality.
It's critical to establish from the start which of the two you are, he continues:
If you oscillate between those two over time, you will confuse not only your own people, but you also confuse the vendors that you work with. If you know where you are, then you can set your strategy accordingly and know which vendors to use for what purposes, and hold true to that route.
SaaS evolution and integration
With Software-as-a-Service (SaaS) becoming pretty much the standard model for delivering HCM software these days, the other thing that's important is to cater for the continued evolution of the various applications you've chosen:
It's not so much about the product you're buying today, it's about how that product evolves over time. You need to create a behavior in your organization where you're constantly keeping up-to-date with what your vendors are doing. It's a bit different from the past, when you bought a system, you'd implement it and leave it for ten years, and then you'd buy the next one.
This is a much more of a fitness exercise where, every quarter or every six months, you look what new capabilities the vendor has. If you take the sunflower as the scenario, you need to be constantly assessing the status of each petal. Is it up to date? Is it healthy? And so on. Sometimes you need to prune them.
With each strategy depending on integrating a portfolio of products together, I wondered what Otter felt about advances in API-based integration, and whether that favors the daisy approach of assembling a portfolio of best-of-breed applications. There are dangers to be aware of, he says, as the imagery takes a sudden and gruesome turn out of the botanical garden onto the veterinarian's surgery table:
Integration is like stitching a tail on a dog. It's really easy to stitch a tail on the dog, if you know what you're doing, because it's very clear which bit's in charge. But if you stitch two dogs together, it doesn't matter how good the stitching process is, it's going to be a mess.
Often that's what happens with organizations. The technical challenges are far smaller today than they used to be, but the governance challenges are the same. The thought I'd always have in your mind as a customer is, what's the dog and what's the tail? When you know that, and when those entities know who's the dog and who's the tail, that integration is very achievable.
Data sovereignty regimes
One complication that Otter suspects will become more prevalent is the need to maintain multiple data stores across different jurisdictions in a global organization. This reverses the existing trend in cloud computing towards maintaining a single global data store, he explains:
For the last 20 years or so, we've made the rather naive assumption that everything's going to continue to globalize. We all store our data in the cloud, it doesn't really matter where the data is. Typically, it's been in the US, but the model is, store the data in this global system suite. It's very efficient, easier to program against. It's much better to have a single data model.
Now that trend is being challenged by the rise in regional data sovereignty regimes — most notably GDPR in Europe, but also evident in Russia, China and elsewhere, he explains:
These countries are realizing that the US has in effect colonized their data, and now there's this pushback — and this is exacerbated by increasing global trade wars. Trade wars historically have been around the price of rice and the price of cars and so on, but it's not inconceivable to see that data flows will be affected by trade wars too. So what this will mean is that organizations will be challenged to store and process data locally.
Renaissance for local HR vendors
Vendors must now have an answer to the new questions being raised by this trend, and it also presents new opportunities to regional and local HR vendors, he believes.
One of the things that we'll need to look at over time is how your vendors are coping with this. Can they offer you distributed physical architectures that have logical consistency? Or are you going to rely more strongly, back to the data model we mentioned earlier, on local petals?
Whereby you take a modern, sophisticated local vendor and you deploy that as your system of record in Russia or in Brisbane — and through better APIs and so on you integrate that into your global systems and your global reporting.
It may well be that the global systems continue to squeeze out the local vendors, but I don't think that the local vendor in HR is dead. I think we'll see a renaissance ... it's not all about just the global vendors, we'll see new local challengers and local players.
When it comes to futures, Otter is refreshingly skeptical of the merits of fashionable tech such as AI and blockchain — on the latter, he comments, "I'm yet to see the awesome revolutionary use case in HR tech." Much of what's being hailed as AI at the moment, he observes, is really just sophisticated analytics based on static algorithms. And the risk of building undetected bias into machine learning systems is very real, he notes. But vendors do have to invest in AI, he says:
I think there's a lot of potential for AI, but I think there's going to be a lot of stumbles and a lot of hype along the way. But I think, as a vendor, if you're not investing significant R&D into getting your AI strategy sorted out, it won't be long until someone's doing it to you.
Innovation and your HR tech strategy
One other area he highlights is innovation in software development processes, which have made it simpler and faster to build new functionality:
You can build great product much quicker and I think this is both important for vendors and it's also important for end users. I think the opportunities for DIY are significantly better than they were a few years ago. You can get your security, you can get your UI controls, either through open source or through your existing platforms.
So you're able to innovate I think much more effectively yourself today than you could have perhaps five or ten years ago. The same innovations that help startups build enable you to innovate yourself as a customer in ways that perhaps a few years ago you couldn't. So there will also be a trend towards DIY.
It's important to understand all of these possibilities, he sums up, but to view them in the context of what's achievable for your own organization. There's one final piece of advice from Otter — don't let vendors dictate your strategy:
Start off by understanding the art of the possible. Spend some time at events like this one we're at today, looking at what the cool kids are doing. Use that to shape your art of the possible, so that when you go out, you actually have some sense of what it is you're looking to do.
But then you need to put that aside and you need to own your strategy. You need to understand what your organization is, what the limitations are, and then you need to build your own individual HR strategy, as your own organization.
A vendor can't build a strategy for you. The vendor strategy, whoever they are, is to sell you more stuff. But you need to own your strategy.