Ellen tried to focus on her job for the last hour of the afternoon by scrolling through the matches in her segment on G-Match. There were so many women looking for men with a bit of something about them, but within their score range. In Ellen’s view, the few men within the segment who were available and in range were all ugly, stupid, congenitally challenged or very dull. Unlike the ads for the service, which showed perfectly healthy, loving couples meeting online and ending up married and cuddling their perfectly formed children, for most people with average PQs, it was compromise all the way. The balance of manageable scores and attraction was the game she helped to manage. Once you get outside your PQ range, there might be good looking, bright, interesting men, who seem to be suitable, but have something lurking in them: a propensity to instability, a dangerous or aberrant streak, or perhaps a biological weakness which would manifest itself in a less than able child, an off-the-scale insurance cost you couldn’t afford.
Ellen’s Insights Analyst job entailed vetting and approving marginal matches in the 26-30 age range in South Central London. A marginal match was classified as outside the normal monthly premium manageable by two working people, or a family. Being just 29, Ellen was managing a segment she understood. this was her own demographic segment, on her home turf, which was why she’d been given it. The whole Insights team tended to work on matches within their own age range or slightly older, and across a geographical area that they would be familiar with, as it produced better results than matching people they might have had less affinity for.
Strike rates were everything in Insights, since it was a very performance-driven department. Insights was always seen as the career stepping stone to Claims Strategy or Deportation, both of which tended to attract the Oxbridge graduates and the high flyers, most of whom were men. To succeed in Zurich, you needed to be cut-throat, connected, hard working and male. Ellen knew she didn’t stand much of a chance of promotion because her PQ demonstrated a softer side to her character, because she didn’t have a daddy on the Board, and because she lacked the requisite tackle. She worked hard because the alternative was joining Jodie on the Happies in Tower Hamlets.
The strike rates for successful matches were compared continuously between human and AI managers, with a view to reducing the company’s dependence on human involvement as soon as AI outperformed people. A success was deemed to be a match which lasted at least three months, comprising at least three dates, and there were bonus points for six and twelve month matches. A marriage resulting from a match earned you more points, but most people on the team were happy getting their three month points. It used to be about first dates, and helping to get new G-match registrations, but that all fell by the wayside as the algorithms became more sophisticated and took into account the truth that fancying someone on a first date was a poor indicator of the chances of a longer term relationship. Everybody knew this to be true, but nobody really understood why.
Ellen was part of the South London team which reported in to Jade, who was one of five Insights Managers responsible for the UK. Jade covered the whole of London and South East Region, using four teams, and was the most senior of the five regional managers, despite having spent a lot less time in the job than the other four. That was partly a testament to her aggressive ambition and diligence, and undoubtedly reflected the power of her father. The 50 analysts in the Insights Department were all on the same deal, and it was one of the most labour intensive departments in Zurich which was why so much work was being put into R&D to improve the AI strike rates and reduce the staffing levels.
Jade led a team of twelve, soon to be six, PQ Analysts. If she was going to make the cut at the end of the month, Ellen would probably be given 30-35s as well, since Zurich preferred to push PQ analysts up the age range, rather than down. If you have to help match younger people, you tend to bring your own history into it, and you make poorer choices because you hold romantic memories of how things might have been, or downhearted memories of how they were. But matching older people is easier. Nobody looks at the future through rose tinted spectacles, after all.
Ellen had always tried not to make biased decisions based on her own taste in men, but to go with the analytics and how well the couples looked together on screen. It was an intuitive process, and she was pretty successful at choosing. If that wasn’t the case, the algorithms would have replaced the whole department long ago, as the computer would have a higher strike rate than its operator. In many ways, analysts had taken a more and more marginal role since the indefinable intuition they brought to matches was the last element yet to be modelled effectively by the systems architects. The algrithm was already able to make initial matches infinitely more quickly, and the basics were always right for first dates. People met and liked what they saw, and got together for dating, as G-Match advertised. But it was later, when the first, biological attraction wore thin that the auto-matches tended to fall down, and the Analyst matches tended to be more resilient. It had been one of the mysteries of G-Match which loads of modelling resources had been poured into. What changes between that attraction, driven by sexual lust and desire, and the feelings which are laid down during the courting process, which leads to stable relationships and families? Building a mathematical model for predicting factors in that change, based on millions of G-match experiences, was the job of Software Development and the Psychodynamics unit, but everyone in Insights had their own theories.
“One of my guys has been on about twenty first dates, and a few second dates, but he finally chose this girl I wouldn’t have expected would be right for him. She scored higher on submissiveness than most of the others, but also had quite high scores for independent decision-making and persuasiveness. She was lower scoring on alure, BMI and problem-solving than almost all the rest. I think he just found her less hard work, and he was tired of dating.” Ellen was musing over the issue with Magda in the coffee room.
“Yes, but I’ve had the opposite too. Remember that life coach who picked the bloke with the anger management issues? They never stopped fighting, and his PQ was way higher than the others we offered her, but once they got to the three month point, his PQ kept going down. She seemed to choose him as a project, and whatever happened, she beat him into shape, which must’ve been what he secretly wanted. God I’d hate to have such a weak man!” Magda had married a strong silent type in the police.
Ellen always felt that her pairings were based on a particular look in the eye of the candidates, something indefinable which put two people together who belonged together. G-Match had millions of successes and failures to compare, and not just the initial data on each, but also the continuous monitoring which a couple provided through their informagear and camfeeds. The G-Match algorithms were self-improving, and the first anniversary strike-rates of computer matches were catching up with the best of the Analysts’ choices. Inevitably, Ellen’s future in the role was going to be short-lived, once the computer model overtook human intervention.
When the announcement was made about the cuts, Jade assured Ellen that she was in the top quartile on strike-rate and that it would stand to her when they made their choice. But Ellen knew how little this platitude was worth, given Jade’s ultimate self-interest and her protected status.