the idea, the index

When you think about a system that does some “matching”, we mostly have some preconceptions about how that might work. RegenCHOICE thinking at its basis, works very differently from other matching systems, including the ones people are mostly familiar with: recruitment and dating. These differences are essential to the functionality as well as the ethics of RegenCHOICE. You can think of these three points that follow as the USPs – the unique “selling” points,¹⁾ even if the service is provided gratis.

Robert kindly put this into ChatGPT to summarise. You can see ChatGPT's shorter versions here.

When we think of two things matching, it is very easy to imagine the two things being the same or very similar, like Snap (card game), or like matching colours. But a moment's thought about dating services shows that is far from always true. If I am looking for a potential partner, the fact that I am a man does not imply that I am looking for another man — more likely, a woman. I may be looking for someone of a similar age, but also I may be looking for someone of a different age. Dating systems always cover these basics first, taking account of the basic asymmetries; but after that, there is a temptation to “match” people based only on similarity, for example of values; priorities; beliefs; attitudes; interests; hobbies; films that you like. Similarity in many of these areas can often make a relationship easier, but it may not be the case.

On the other hand, if we look at recruitment (from the employer side) or job search (from the side of the job seeker), the picture looks different. A person can't closely resemble a company. So what the company usually does is to create an image of the kind of person who they think would suit the job role they are advertising — then, some kind of similarity metric can be used.

While, for dating, it's always the preferences of both sides that matter, for employment, what about the preferences of the job seeker? A brief moment's reflection will confirm that the recruitment process is highly asymmetrical. The dice are loaded in favour of the employer — this is hardly surprising, as it is they who pay for the service. The person seeking a job generally has to rely on a bare minimum of information about the job and the company. Pay: yes; location: yes; those are givens. But many of the other ways the job seeker typically searches are around the similarity of themselves to what is described in the job advert. It's less often about the company, or the work environment, because these are often not advertised directly. To find that out, the job seeker has to do their own time-consuming research, or just wait to ask questions in an interview, after they have already gone to the trouble of applying for the vacancy.

What if, instead, it was arranged like this … the company specifies the characteristics of the person they want to employ, and the individual specifies the characteristics of the job and the company they want to work for. Then, the “similarity” matching is truly two-way – or four-way, as a colleague once described it – in two distinct and separate halves. What the company wants is matched against what the individual answers, and at the same time, what the individual wants is matched against what the company answers. This is the essence of two-way matching, and it is not seriously done in any online recruitment service that I know of.

But if we did have this kind of two-way recruitment service, consider what the consequences might be. Individuals would only find potential jobs where their requirements were met. Employers would, as now, find individuals who met their criteria; but more significantly, they would only see the individuals who would fit best into their company according to the criteria of those individuals themselves. The individuals wouldn't have to wade through so many applications, and they would be much more likely to find a post where they were really happy and engaged — assuming that they knew what would really satisfy them. Think of how much time, effort and money would be saved if companies could recruit only those employees who really wanted to work in the kind of environment they offered. We would obviously expect that their staff turnover would greatly reduce, the engagement and productivity would increase.

A win-win if ever there was one.

And of course this approach doesn't only apply to employment, but to any situation where an individual is looking for an opportunity with a group or organisation, and the organisation or group is looking for an individual to join them. Think of a project, or a co-living situation, as examples that have been alive for me. No doubt there are many more examples — let me know!

Online dating sites have been around for many years now. Before that, apart from the so-often-advised “get out there and meet people”, and going through one's old-style social network (not the online variety), the other way was personal “lonely hearts” ads in newspapers and periodicals.²⁾ How can you make yourself appear attractive in a few words? How many people would you have to contact to have a reasonable chance of finding someone with whom you really fit well?

Now I've noticed over the years that there seem to have been two distinct approaches to online dating (and just before that, “computer dating” by post.) On the one hand, you say what you are looking for, along a few basic dimensions, and answer the same questions yourself. Doing a two-way match for a few standard simple criteria isn't hard, even for a computer. But again, based on these few criteria, how many people will you have to wade through to find someone more than superficially compatible? Of course, in any such system there will be a few lucky ones, and they will be the ones used in the promotional material for the service. The less lucky majority will need to extend their subscriptions — no bad thing for the finances of the business that runs that service.

On the other hand, there are, and have been, services where you are expected to answer a whole lot of questions. Yes, this is more likely to cover areas that are important to you. But, and it is a big reservation, who has both the time and the inclination to fill out such long, long forms? The people who meet each other in this way will be pre-selected on their willingness to fill in long forms. Which at least is rather more fair than the highly priced services, where you are more likely to find someone else who is much better-off than average, if that's what you want. And even then, how likely is it that the people running the service have actually included the questions that are really important to you? How long is the “long tail” of questions that matter to individuals? I'd guess, pretty long, so likely many questions important to people still are not there.

How RegenCHOICE gets round this challenge is another of its central points. In practice, people won't be having very many specific requirements — if they did, they would be unlikely to find anyone at all, outside their dreams. And the only questions you need to answer are the questions asked by the people who you are interested in — in other words, those people who meet your requirements. As you answer each of these questions, the field quickly clears, and the people who are looking for someone who is not you drop out of the picture. And, naturally, RegenCHOICE is designed to feed to you just these questions that you need to answer, to determine whether there is a match or not.

This is the key point here, then: that there can be as many questions as are wanted, and still only a small fraction will need to be answered in any “enquiry” (as I have called them). It's not simply the time-efficiency by itself that matters here. It is that people tire, and so often settle for jobs, partners, etc. who are not really what they want, or could get if they persisted. Here with RegenCHOICE the persistence needed to get a really good fit is minimal, hopefully within reach of all.

One other important feature emerges from the first two, and it promises to make a real difference in setting RegenCHOICE apart from other services.

Put together these two considerations. First, what really matters most to people can in some cases be quite sensitive, private even. Second, I've confirmed in recent conversations the observation that many people – probably most people – can be reluctant to put their normal, public details in any open place on the web, let alone sensitive information. Putting those together, if we are talking about any but the most extravert individuals, it is likely to take a lot of trust and persuasion to persuade people to enter their deepest desires onto any computer system, unless then can be assured that only the people they want to see it will see it.

Many systems understandably are either shy of very personal information, or can't cope with the large number of questions that it would take to have all those questions written in to the system. RegenCHOICE is able, as noted in the previous feature, to deal with large numbers of questions, so there is room for the sensitive ones as well as the everyday social or public answers. And then, the two-way process described in the first feature ensures that, if RegenCHOICE is used as envisaged, the only people who even get close to seeing that highly personal information are the ones who satisfy the conditions laid down by the owner of that information.

If you are asked a personal question, as part of the RegenCHOICE process, you can naturally decline to answer it. This will mean that you won't match anyone who requires an answer to that question. Alternatively, you can ask equally personal questions of the other, so limiting the people who even get to see your anonymous answers. The way the system is currently envisaged, both people will need to approve of the anonymised answers of the other before connecting by exchanging some contact details, and I guess that will be enough basic security to reassure people that it will be as safe as can be to ask for and answer the questions that really matter.

As a result of being able to ask those deeper questions, there is likely to be much less time and energy spent worrying about when or how to bring up potentially awkward but important facts. And so, the promise is of much higher quality matching than is possible with other systems.

• RegenCHOICE