The Kitchen Whippy Project: Three Gadgets, One Flake, No Van
There is a particular swagger to a van-made 99. The cone, the curl, the flake wedged in at an angle that suggests confidence. In November 2026, with our local van hibernating and the freezer full of half-bags of peas, we wondered whether a normal UK kitchen could fake it. Not improve it. Fake it. So we cleared the counter, consulted three recipes of varying dubiousness, and invited the family to judge the results with the ruthless honesty usually reserved for Sunday roast potatoes.
The rules: same flake, same family, same unreasonable expectations
We set four tests: texture (soft-serve smooth or icy?), cost (supermarket prices, not restaurant prices), mess (counters, floor, hair), and the Flake Test. The Flake Test is simple: insert a standard Cadbury Flake and see whether it stands upright for 30 seconds or collapses into sad chocolate rubble. Each method had to produce six child-sized portions. We used a £1.20 supermarket own-brand vanilla extract, double cream, condensed milk, whole milk and a bag of flakes. No stabilisers, no specialist machine, no pretending a food processor is a Pacojet.
Method one: the no-churn condensed-milk base
This is the internet’s favourite shortcut, and for good reason: it is almost impossible to ruin. We whisked 600ml double cream to soft peaks, folded in a 397g tin of condensed milk and a generous glug of vanilla, scraped the lot into a lined loaf tin and froze it for six hours.
The result was rich, dense and scoopable, with a texture somewhere between good ice cream and a very posh clotted-cream experiment. It did not look whippy. It looked like ice cream that had been told to relax. Texture scored 7/10. Cost came in at about £3.20 for six portions. Mess was low, unless you count the sticky tin lid and the child who volunteered to lick the spatula and then vanished. The Flake Test was excellent: the flake sat on top like a flag on a hill, because the base was firm enough to hold it.
Useful tip: stir the mixture once after about two hours to break up icy edges, and use a warm spoon for serving. Downside: it needs planning, and you are left with half a tin of condensed milk, which is less a problem and more a dessert waiting to happen.
Method two: the piping-bag swirl
Emboldened by the first result, we chilled a second batch for 90 minutes, loaded it into a piping bag with a star nozzle and piped rosettes onto a chilled tray. Ambition was high. Dignity was low. The first swirl looked like a seaside snail. The second looked like a soft-serve that had seen things. The third was acceptable if you squinted and did not compare it to the photo in our heads.
After 45 minutes in the freezer, the swirls held their shape but tasted airy at the edges and solid in the middle. Texture scored 5/10. Cost was about £3.60 for six, including a new piping bag but not the nozzle we already owned. Mess scored 8/10: cream on the cupboard, a clogged nozzle, and a family debate about whether licking the inside of the bag counts as hygiene or initiative. The Flake Test earned 6/10. If you insert the flake immediately, it stays upright. Wait a minute and it leans like a lamppost in a storm.
Useful tip: chill the tray first, do not overfill the bag, and pipe close to the surface rather than trying to draw a perfect curl from a height. It is a good birthday-tea trick, not a route to perfection.
Method three: the blender trick
This was the wildcard. We froze 300ml whole milk and 100ml double cream in ice-cube trays overnight, then blitzed the frozen cubes with two tablespoons of sugar, vanilla and a splash of milk until smooth. The result was genuinely soft-serve-ish: glossy, pale and instant. Texture scored 8/10. Cost was around £1.60 for six small portions. Mess scored 9/10, thanks to the blender jug, the lid, the spoon, the counter and one unexplained splash on the wall that we are still discussing.
The Flake Test was a 4/10. The flake sank like a tiny chocolate Titanic. This method melts fast, so it is a serve-immediately situation. Useful tip: chill your bowls, blend in short bursts, scrape down the sides, and hand out spoons before you start blitzing. A pinch of xanthan gum helps if you have it, but it is not essential for a one-off kitchen caper.
The scorecard
| Method | Texture | Cost for six | Mess | Flake survival | Faff |
|---|---|---|---|---|---|
| No-churn condensed milk | 7/10 | About £3.20 | Low | 9/10 | Medium |
| Piping-bag swirl | 5/10 | About £3.60 | High | 6/10 | High |
| Blender trick | 8/10 | About £1.60 | Very high | 4/10 | Low to medium |
What actually worked
The blender trick won on texture and speed, but only if you serve it instantly and accept that the flake is decorative rather than structural. The no-churn base won on reliability and flake-holding, and it is the one we would make again for a planned family tea. The piping-bag swirl won on nothing except comedy and the fact that the children found it hilarious. It is worth doing once for the photos and the lessons in humility.
None of them tasted exactly like van soft-serve, because van soft-serve is mostly air, sugar and mystery, and our kitchen is not equipped to manufacture mystery. But all three were eaten without complaint, which in a family of six is practically a Michelin star.
So, is home whippy worth the faff?
Compared with the van, a home batch is cheaper. Six decent 99s from a van might set you back £15 to £21 depending on your postcode and the server’s generosity with the flake in 2026. Our most expensive method cost £3.60 for six portions, before you count electricity and washing up. But the van also includes an outing, a conversation with a stranger, and someone else’s flake placement. You cannot blender that.
Our verdict: home whippy is worth the faff if the faff is the point. For a quick Tuesday treat, no. For a rainy Sunday when the children need a project and you need to feel briefly heroic, yes. Fake it till you whippy, by all means. Just do not expect the van to start worrying.
