Benchmarking a generalist role is hard enough. Benchmarking a niche one – or a role that barely existed two years ago – is a different challenge entirely.
The data is thinner. The peer groups are harder to define. And whatever methodology you land on, you still need to defend it internally and make sure it holds up next to the rest of your framework.
Sunny and Alistair have both had to work through this in practice – navigating limited market data, emerging skill sets that surveys haven't caught up with yet, and the pressure to maintain consistency across a compensation structure that wasn't built with edge cases in mind.
This session covered:
As always, there will be plenty of opportunity to ask questions, share experiences, and learn from fellow reward professionals who are navigating the same challenges.
If you're more of a reader than a watcher, here are a few of the most interesting insights from Sunny and Alistair's discussion on benchmarking niche and emerging roles.
Before anything else, it's worth separating the two.
Niche roles are ones that have existed for a long time – baggage handlers, aircraft leasing specialists, clinical technicians – but they're specific to particular industries and so don't appear reliably in standard survey data.
Emerging roles are new, often moving fast, and driven by technology shifts: AI engineers, prompt engineers, EV battery specialists, rocket propulsion engineers.
Both present data challenges, but the way you solve them differs.
For niche roles, there's often a club survey route – gathering a peer group of companies with the same roles and pooling data in a confidential, structured way. That takes time and budget, but it builds the kind of internal credibility that's hard to get any other way.
For emerging roles, that approach rarely works. The peer group doesn't exist yet, the data is too thin on the ground, and there usually isn't time. You need to move faster with less.
When the survey data isn't there, the natural reaction is to go looking for more sources.
But, when you strip it back, often what's actually missing is a clear enough picture of the role itself. What problem does it solve? What skills and competencies does it actually require? How does it map against roles that already exist internally?
Both Sunny and Alistair instead first aim to better understand the role first, by speaking directly to the hiring manager, then level it internally against the job architecture, and only then go to external sources.
Getting under the hood of the role first means the external search is more targeted and the eventual number is easier to defend.
💡 Practical application: Before opening any survey tool for a role you’re struggling to find data on, ask whether you have a comparable role internally. Even a 70-80% match can give you a defensible anchor point – and a justified premium on top of it.
When the data is thin, the credibility of your process has to carry extra weight.
That means keeping a record of every source used, why it was used, who provided the information, and what confidence level you assigned to it. Recruiter conversations, job adverts, bespoke reports, candidate expectations – all of it needs to be logged.
💡 Practical application: Don't rely on URLs to job adverts. They disappear. Screenshot them and save them to your evidence file. And when you only have data for two or three levels of a role but need to fill out a full framework, document the regression logic you used to fill the gaps – because six months later, nobody will remember where the numbers came from, including you.
Employees are increasingly arriving at conversations with AI-generated benchmarks that appear authoritative. Both Sunny and Alistair were clear on this: discount it entirely.
The same prompt given to the same tool ten times can produce ten different numbers. The data skews heavily towards the US because there’s an abundance of data there. It has no awareness of your pay philosophy, your peer group, your equity offering, or the strategic decisions your executive team made when they defined your compensation position.
The more useful pushback, especially with senior stakeholders is to remind them that your pay position was set in the room with them. The scope decisions – which company sizes to compare against, which markets, which roles – are what make the data meaningful. An AI engine doesn't know any of that, and the number it produces reflects none of it.
💡 Practical application: Build a short section into your compensation philosophy and manager training that addresses AI-generated data directly. Having a clear, consistent answer ready means you're not caught off guard when a candidate or hiring manager brings one in.
Your monthly dose of market insights and expert perspectives
Join Henriette Zimmer and Gina Trautwein for an honest conversation about retention – spotting the risk before the resignation letter lands, deciding which lever actually works, and knowing when pay alone won't cut it.
Leapsome doesn't negotiate offers. Jessica Zwaan, VP People Strategy and Operations at Leapsome, explains the philosophy behind that decision, and Natasha Delisle-Barrow, CPO at Ravio, brings the market data on what negotiation does to pay equity. Expect a practical case for calculated pay, and what to do about the gaps negotiation has already created.
Join Ekaterina Potter (Global Director of Reward and Belonging at commercetools) and Daniela Černá (Global Compensation and Benefits Lead at Aeven) for an honest conversation about running merit cycles in practice – the wins, the parts that broke, and what they'd do differently.
