iDiet — why people buy, and what the site is throwing away
Deep dive + conversion brief · Dubai AI House
---
1. What they actually are
- **HealthBox FZ-LLC**, Dubai. iDiet is the consumer brand; the kitchen operates as
Lifestyle Diets Catering. Roots in Beirut — the food is Levantine-Mediterranean, not
gym-food.
- **Clinically supervised.** Registered dietitians (Sabine Karam among them) sign off the
menus. Macros are printed on the box itself.
- **The app is the business.** iOS since Oct 2021, currently v2.3, developer HealthBox.
Subscribers pick their dishes in advance, exclude allergens, pause, and swap. Version
notes show a redesigned home page, added "view previous selections," and new payment
methods — this is a product they keep investing in.
- **Logistics.** Cooler bag at the door **before 5am**, across Dubai, Abu Dhabi and
Sharjah. 48-hour window to pause or change.
- **Reputation.** 5.0 from ~499 Google reviews. But only **18 App Store ratings** — the
app works, almost nobody has been asked to rate it.
- **Price band.** ~AED 86–143 per day depending on plan and board; iDetox Full Board at
AED 143/day (AED 2,002 / 14 days, AED 4,004 / 28) is the published anchor.
2. The moat, stated plainly
Everyone in this market delivers calorie-counted food. What iDiet has that Kcal, Delicut,
FITT and Right Bite mostly don't:
1. **You choose your dishes, daily, before they're cooked** — not a fixed rotation.
2. **The food is genuinely good and genuinely regional** — sayadiyeh, tabbouleh, moutabal.
Reviewers talk about *looking forward to the box*. That is the whole retention engine.
3. **Sauce on the side. Allergens removed. Portions signed by a dietitian.**
Read their reviews and the pattern is identical every time: *the food tastes amazing, and
choosing it in the app is easy.* Nobody buys a diet plan. They buy dinner they'll enjoy
and won't have to think about.
3. What the website does with that moat
Nothing. It is a plan catalogue.
| Gap | Cost |
|---|---|
| **No food visible before you pay** | The single strongest asset is hidden behind signup |
| No menu, no dish photos, no macros | Buyer can't answer "will I like eating this?" |
| 12 plan tiles as the entry point | Choice paralysis at the exact moment of highest intent |
| Client-rendered, no SSR | "meal plan dubai" traffic can't find them; Google sees an empty shell |
| English only | Half the emirate reads Arabic first |
| No trial, no 1-day entry | AED 2,000 is the smallest yes available |
| 499 Google reviews nowhere on site | Their best proof is on someone else's property |
| No urgency, no cutoff, no delivery promise on the page | "Before 5am" is a killer line and it isn't there |
The paid click that brought you here — Google Ads, `gad_campaignid=22329376531` — lands on
a page with none of this. That is money going into a page that cannot close.
4. The neuro logic of the rebuild
A Dubai buyer opens this at 11pm on a phone, tired, having decided *something* has to
change tomorrow. Five things decide whether they pay:
1. **Appetite before argument.** Visceral response beats reasoning. The first thing on the
screen is food they could be eating tomorrow morning — not a value proposition.
2. **Endowment.** Let them swap dishes *before* they pay. The moment they change lunch,
it stops being iDiet's box and becomes theirs. People do not abandon things they own.
This is the single highest-leverage change on the page.
3. **Reduce the choice, not the options.** 12 plans is a wall. One sentence about
themselves, and a dietitian rebuilds the box in front of them. The options remain, but
the decision has been made for them.
4. **Real urgency, honestly used.** There is a genuine cutoff — the kitchen cooks tonight.
A live clock to 6pm is not a fake countdown, it's operations. It converts because it's
true.
5. **Anchor small.** AED 4,004 is the number on their site today. AED 29 a meal, versus a
AED 85 DIFC lunch nobody counts for you, is the same money and a completely different
decision. And the first yes should cost AED 99, not AED 2,002.
Proof goes *at* the decision, not the top: 499 reviews, macros on the box, pause any day —
placed immediately above the button, where doubt actually happens.
5. The screen flow
```
1 TOMORROW'S BOX four real dishes, photos, macros, live cutoff clock
↓ tap any meal
2 SWAP SHEET alternates for that slot, "chosen 1,204× this month"
↓ (or) one sentence about yourself
3 DIETITIAN REBUILD Claude picks the plan, rewrites all four cards, flags
pregnancy / diabetes / ED / under-18 to a human call
↓
4 CLOSE Tomorrow only AED 99 · 14 days · 28 days
per-meal anchor, dietitian call promise, pause policy
```
Everything above the fold on a 390px screen is food. The plan names are a filter strip,
not a gate.
6. What I'd say in the room
Hand them the phone. Let them swap tomorrow's lunch themselves. Then say:
> "You already built the thing that keeps people — the app where they pick their food.
> You just don't let anyone see it until after they've paid AED 2,000. Every ad click you
> buy lands on a page that hides your best asset. Let them build tomorrow's box before
> they pay, and charge AED 99 for the first one."
Then the second, quieter point: they have 499 five-star Google reviews and 18 App Store
ratings. There is a whole retention and referral layer sitting unbuilt on top of a product
customers already love.
7. Build
Next.js on Vercel (SSR the menu — that's the SEO unlock), Supabase for menu, boxes and
subscriptions, Claude for the dietitian match, Arabic from day one, WhatsApp for the
nightly "your box closes in an hour" nudge. The existing Java app stays as the ops
backbone; this is the acquisition surface in front of it.
---
*Sources: idiet.ae, App Store listing (HealthBox FZ-LLC), HealthyInUAE field review,
WhatsOn and The National roundups, public DNS/header records. Dish names and order counts
in the demo are illustrative.*