Greetings. 

By the time this lands I’ll have had the house to myself for twelve days, and I want to be honest with you about where that leaves a man. The first few days you enjoy the quiet. By day six you’re narrating your own dinner. By day nine you’ve started reading company accounts aloud to the cat, with feeling, and being mildly hurt when he doesn’t react to the good bits. 

(If it were up to me, Theo would’ve been called Bruce, after The Boss, but I was outvoted by the family. That’s your lang cat cat fact – LCCF. Should every TCWU have an LCCF? I think it should. What do you think Theo? Miaow.) 

Anyway, Theo and I have been going through a set of accounts this week, and they’re a thing of beauty, in the way that a slow-motion replay of someone walking into a lamppost is a thing of beauty. 

They belong to a company called J.P. Morgan Personal Investing Limited. You’ll know it by its old name, because until last October it was Nutmeg, the brand JP Morgan bought for around £700m in 2021, then quietly retired to consumers in November. 

On the face of it – wayhey! – a profit. £12.4m for the year, against a £40m loss the year before. Fourteen years in the red, and finally, some daylight. High fives and slow-motion montages all round. 

Now look one line up. 

The loss before tax was £58.2m, worse than last year, by eighteen million quid. On a trading basis this is the worst year the business has ever had. What turned that £58m crater into a £12m profit was a tax credit of £70.6m, and once you get into note 12, it’s magnificent. 

Here’s what happened, as far as Theo and I can tell. For fourteen years, Nutmeg’s losses were worthless on paper, because you can only bank the tax value of a loss if you expect to make enough profit to use it against, and Nutmeg never did. Fair enough. But now it’s owned by one of the most profitable banks on earth, so those losses can finally be set against JP Morgan’s tax bill instead. The group takes the losses off Nutmeg’s hands, for cash (there’s a £46.5m receivable for exactly this), and a big hairy deferred tax asset gets recognised for the first time. The result: a £70.6m credit, most of it prior-year adjustments, none of it a single new customer walking through the door. 

Strip the wizardry out and here’s the actual business. Revenue grew nicely, fee income up 23% to £48.5m. Genuinely good. But costs sprinted straight past it: marketing up 42% to £18.2m (in the very year they announced they were killing the brand it was paying for), the intercompany charges from JP Morgan nearly tripling to £19m, headcount up by a sixth. Assets under management now stand at £9.4bn. And it lost more money than ever. The bigger it gets, the more it bleeds. The scale economics everyone was promised are, at £9.4bn, still not visibly in the building. 

And the maths on why is brutal. Fee income of £48.5m on £9.4bn of assets is a yield of about 50bps – which is a perfectly normal number for a robo, and that’s rather the problem. At that rate, every billion of assets brings in around £5m. Set that against a £108m cost base and you need north of £20bn under management just to break even. They’re at £9.4bn. So, the honest answer to “when does this make money” is “at roughly double the size, assuming nothing else goes up” – and the trouble is everything else keeps going up. 

Now, I want to be fair here, because it would be asy to read all that as “Nutmeg was rubbish.” It wasn’t. The tech is genuinely good, always has been. The app’s clean, the onboarding’s smooth, the proposition does what it says. If the product were the problem, none of this would be interesting. That’s exactly the point. You can build a lovely thing, run it beautifully, get famous doing it, and still not make the sums work, because the cost of getting strangers through the door outruns what they’re worth once they’re inside. Nutmeg spent fourteen years, north of £200m of accumulated losses, and £180m of fresh capital from its parent in the last two years alone, proving that the hard part was never the tech. It was the getting-and-keeping of customers, profitably, before the meter ran out. 

Here’s the bit that might make JP Morgan twitchy. There may be another monster coming over the hill. 

Targeted support went live in April. It’s aimed squarely at the advice gap, and the firms best placed to use it are the ones that already have millions of trusting customers sitting in an app: the big banks, the pension providers, the household names. In other words, exactly the brands consumers already trust with their money by default, now able to gently steer that money towards their own investment products. 

Think about what that does to the economics we’ve just been staring at. Nutmeg’s whole existence was a fourteen-year, £380m-and-counting exercise in manufacturing trust from a standing start. Targeted support hands a shortcut to everyone who already has it. If your bank can quietly suggest its own ISA to you inside the app you already open every day, the cost of acquiring you drops to roughly nothing, and the plucky challenger who had to buy your attention one tube passenger at a time is suddenly racing people who got the customer for free. JP Morgan, to be fair, is now one of those people. But it paid £700m and a brand to get there. 

The directors, for what it’s worth, note that they’re “comfortable that the Company’s financial performance aligns with its strategy.” I’d love to be that comfortable about anything. 

Right, that’s your lot. Theo’s unmoved, as ever. He’d have been an excellent Bruce. 

I have to go with Bruce for the music this week. I asked Theo what I should pick. He said “Thunder Road. Miaow” I said, isn’t that too obvious? He said “what, even the Hammersmith Odeon 1975 one, that’s iconic. Miaow” I said, fair point but could we go off the beaten track a bit. “How about his cover of Royals by Lorde? That’s a deep cut. Miaow”.

Good choice. Told you he’d have been an excellent Bruce.