Like every time after a visite to HR Tech World I write a review of the most interesting start-ups I’ve seen at the event. These are the previous articles: november 2015, march 2016 and november 2016. I never repeat a start-up to watch in these articles, if I meet a company that’s developed they get a separate article like artic shores got last year for example.
Let me start by saying I was unpleasantly surprised by the lack of some start-ups I did not see. Hence, no follow up articles. None of the start-ups to watch from november were present in London, so no updates. I met some interesting ones, although I must admit the level of start-ups exhibiting wasn’t the best quality I’ve seen at HR Tech.
Seed Link
An interesting company not exhibiting, but presenting, was Seed Link. Since they did not exhibit I can’t say I’ve seen a demo and usually I don’t write about companies I haven’t seen ‘in action’ yet, however, the product promise is pretty good and since they had a client with them presenting, I have a tendency to trust them. Seed Link uses language technology to identify markers in resume’s and cover letters telling you more about the personality of a the candidate. Although their claims go beyond that (matching with organisational culture), that alone is enough for me to be interested. No separate assessment, but assessing someone’s normal language in resume’s, cover letters and potentially other questions you ask during a digital interview. In many (or all) languages. Since they also have Dutch clients using the technology, it’s not just English (or Chinese) based. I will be keeping a very close eye on them this year.
Devscore.io
Devscore is a start-up that helps you assess IT applicants. Before you think, yet another? Didn’t we go trough this before (with IT test online in march 2016 for example)? They are different. Different in this case has an upside and a downside. Basically what they do is they give you a rating of any applying developer per language / framework. That’s good. It’s not in general code, it’s for the code you want him or her to develop in. The way they do it differs a lot from the existing tools, like Codility and IT test online. They look at code developers actually made. So they look at their github repository or a developer can actually upload a piece of code he or she has built already. That saves time and effort on the side of the applicant. One of the main reasons not to use Codility or IT test online, applicants don’t always like it. They did do well to make sure code wasn’t copied and it’s original, so fail saves are built in. However, they can’t test other parts of programming like speed and consistency. If someone wrote code, you have no idea if it took him an hour or a day. With IT test online you do. You also don’t know if a developer is consistent and builds in the style of existing code so it’s ‘readable’ for others. Also something you can test with IT test online. So the upside of this: it costs developers a lot less time to apply. The downside: you get less information about the developer. It is however an interesting new addition to the IT testing applicants market.
Predictive Hire
Predictive Hire is an Australian start-up that make huge promises and uses a lot of hot marketing lingo, but in between they actually seem to have a pretty decent product too. Basically what they do is they look at your company’s KPI’s and most important performances. They then look at your best employees on these KPI’s and get them to fill out a questionnaire. Next step they let all of you applicants fill out a short version of the questionnaire, since many data points have no predictive value anyway, and you get better matching. They system is unique for your company and applying machine learning the algorithm will improve over time. I like the basic idea, but I’m still in doubt about the execution. The part I’m no fan of is the questionnaire. Questions, next to being easily manipulated, can only show you so much. Important features like ‘acting under stress’ or ‘anticipation’, ‘memory’ and ‘reaction speed’ cannot be tested with a questionnaire but could for many jobs be major indicators. Gaming could be a solution for this. So I will be watching them, since they believe in company specific algorithms. Next time I meet them I hope to see more psychometric datapoints in their algorithm, based on gaming or other ‘new’ forms of testing.
Pinsight
Pinsight is a compamy to watch, not because I was immediately sold on the concept, but because it’s easily sellable (it’s close to what we know) and from there they might be able to upgrade the product. What they do is they brought the old fashioned assessment with role playing to the webcam. And of course they built a system around it with more interaction like you would have when spending a day in an assessment centre. So during the course of a week a candidate would get a few e-mails he or she needs to interact with and a webcam interview with an actor. The actor part, with much human interaction, makes it more expensive and less scalable. With all the modern AI stuff ging on, I assume this will be replaced within the year with a virtual assessor. I wrote about Intervyo in October of last year, they should be looking at cooperating with them I guess. The part I’m also missing right now is the modern ways of gathering more data. Like Intervyo (and others) you can use micro expressions and natural language processing to get more information in. For now, the product in my opinion offers too little added value beyond what’s already out there. They digitalised an existing process. Now it’s time to add to the process. If they can do that, it’s an interesting start-up to watch.
Pymetrics
Pymetrics is a game based assessment. I love those. I’ve written about Artic Shores in the past and I’m a fan of them as well. Pymetrics told me they get to the same thing, but from different scientific backgrounds. Pymetrics comes from a neurological background, while Artic Shores uses more applied psychology. Since I have no idea about the major differences in outcome (I get the basic difference in the academic fields, just not how the outcomes in a game differ), I can’t judge the validity of one or the other. I do see a major difference in their business model. Pymetrics seem to focus mainly on getting candidates to play the games and match them more with a jobboard style system to clients. Artic Shores focusses more on matching candidates that already apply at your company to the best performers within your company. So for now, they don’t really seem to compete, although I assume Pymetrics will more to the ‘test your applicants’ business as well. While Artic Shores is a UK company, Pymetrics is an American one and it shows in their communication and market focus. They might both get a piece of the pie, with room for other players as well.