Fifty Restaurants and Still Hungry

Last Friday I spent 38 minutes picking a place for dinner. I opened Google Maps, sorted by rating, scrolled through 50 spots in a 2km radius. Then I cross-checked the top eight on TripAdvisor. Then I read recent reviews on the ones that survived because a 4.7 average from 2019 means nothing. Then my wife asked if I'd just picked somewhere and I admitted I hadn't.

We ate at the place two minutes from our apartment. The one I could have walked to before opening any app.

The platforms don't want you to decide

Yelp, Google, TripAdvisor — they all make money the same way. More listings means more ad inventory. More reviews means more SEO surface area. More results means more time on platform. Their incentive is to give you a list, not an answer.

A list looks fair. A list looks democratic. A list also means the platform never has to be wrong, because the decision is yours. If you pick spot number seven and it's bad, that's on you. You should have read more reviews.

So they show you 50 results when you wanted one. They sort by "relevance," which is a word that means whatever the algorithm wants it to mean this quarter. The top three are usually sponsored. Spots four through fifteen are juggling for position based on how recently someone reviewed them and whether the owner replied. The actual best restaurant in your neighborhood might be at position 23 because the chef doesn't run a marketing operation.

Reviews stopped being signal

Reviews used to work. In 2010 if a place had a 4.5 with 200 reviews, that meant something. Now every restaurant has 4.5 with 2,000 reviews because:

• Owners ask every happy customer to leave one
• Unhappy customers mostly don't bother
• Review-for-discount schemes inflate the average
• The 1-star reviews are usually about parking or a rude waiter on a Tuesday

The information density of a review average has collapsed. A 4.6 and a 4.2 are statistically the same restaurant. So you start reading individual reviews, and now you're 22 minutes deep into someone's complaint about the bread basket from eight months ago.

The cost is real

I don't think people talk enough about how exhausting this is. You wanted dinner. You ended up doing research. The decision fatigue is the whole reason food delivery apps win — not because the food is better, but because the choice architecture is smaller and the photos are bigger.

Picking a hotel for a weekend away is worse. Picking a contractor is worse than that. Every consumer search has turned into a part-time job, and the platforms benefit from the work taking longer.

One answer instead of a list

I built Zerch because I wanted the opposite shape. You ask for the best Italian in Marina, or the best dentist in Jumeirah, or the best electrician in your postcode. You get one answer. Not a top ten. Not a sponsored carousel. One.

You can disagree with the pick. You can ask for the next one. But the default is a decision, not a menu.

This means Zerch has to be wrong sometimes in a visible way. A list is never wrong because it never commits. A single pick can be argued with. I think that's a feature. The platforms that show you 50 options are protecting themselves from criticism. They've also stopped being useful.

The promise is in the tagline: we picked the best so you don't have to. The 38 minutes I lost last Friday is the problem I'm trying to delete. If you've had your own version of that Friday, zerch.me is the bet I'm making on what comes after the list.