Analytics for your app: the few numbers worth tracking

What numbers should you track in your app?

The short answer: analytics for your app should answer just six questions at the start. Where do people come from? How many tried the core thing your app does? How many paid or booked? How many came back? How much money actually came in? And how many stopped or cancelled? Every other number is either a detail of these questions or a vanity figure that looks good in a slide and changes no decision.

Many owners open an analytics tool full of charts and close it none the wiser. The problem is rarely too little data; it is too much. A perfume store, a restaurant or a clinic needs to know where customers leak between the ad and the payment, not memorize twenty acronyms.

Below: the few numbers worth tracking, the events behind them, an in-app dashboard, reading seasons honestly, and a weekly decision routine.

A number that never changes a decision does not deserve a place on your dashboard.

The six core metrics and how to calculate them

These are the metrics we recommend for any booking app, store or service in its early stage. Call them your KPIs if you like; what matters is knowing exactly how each one is calculated.

Conversion rate is usually the most important number for a store, and churn is the one that matters most for subscription apps such as gyms and courses. Activation, often forgotten, shows whether visitors understood your app at all.

Define events before you build dashboards

An event is something a user did that you want to count: sign-up, add to cart, start checkout, successful payment. A dashboard is only a way of displaying events, so if the events are vague or duplicated, every chart is wrong no matter how polished it looks.

A simple event list that fits most apps

• signup_completed: finished sign-up and verified their phone number.
• item_viewed: opened a product or service page.
• cart_added or slot_selected: added to cart or chose an appointment.
• checkout_started: reached the payment page.
• payment_succeeded and payment_failed: the payment outcome, with the method (mada, Apple Pay, Tabby or Tamara).
• order_cancelled or subscription_cancelled: with the reason, if the user picked one.

Use consistent English names and attach a few useful properties to each event: channel, language (Arabic or English), device type, and city at a general level. Do not attach names, phone numbers or addresses; personal data has no place in analytics.

Know where every riyal comes from

Give each campaign and channel its own tagged link: one for the Snapchat ad, one for TikTok, one for your Instagram bio, one your team shares on WhatsApp. That way you see which channel brings paid orders, not just visits. A channel with lots of visits and few orders may be your most expensive one.

Build the dashboard inside your own app

External analytics tools are useful for traffic, but the real money numbers live in your own data: orders, payments, bookings and subscriptions. So the best dashboard for a small business is often a page inside your admin panel that reads straight from your database. For the details, see build an admin dashboard for your app.

1. 1
   Write the six questions

   Phrase each metric as a question about your business: how many paid orders came from Snapchat this week?

2. 2
   Pick one source per number

   Does it come from the orders table, app events or the payment gateway? One source per number prevents conflicting figures.

3. 3
   Ask the agent for the dashboard

   Describe the exact cards, charts and filters you want, with an adjustable date range and Hijri plus Gregorian dates.

4. 4
   Reconcile once by hand

   Compare a week of sales on the dashboard with your gateway report. If they differ, find out why before you trust it.

5. 5
   Set access rights

   The accountant sees revenue, support staff see orders, and not everyone sees everything.

How do you read numbers during Ramadan, Eid and other seasons?

The Gulf market is clearly seasonal. During Ramadan, peak hours move to after iftar and close to suhoor; before Eid, gifts, perfume and clothing spike; then things go quiet during Eid itself. National Day, White Friday, Eid Al Etihad and the Dubai Shopping Festival all create jumps that do not mean your app improved.

• Compare season with season: this Ramadan with last Ramadan, not with Sha'ban. Ramadan shifts about 11 days earlier each Gregorian year, so comparing by Hijri date is more accurate.
• Mind the weekend: Friday–Saturday in Saudi Arabia, Saturday–Sunday in most of the UAE. Compare full weeks with full weeks.
• Separate the discount effect: more orders during a big sale do not prove your product page got better.
• Watch the hour of day: if most orders arrive after taraweeh, do not schedule app updates or WhatsApp campaigns then.

Analytics and personal data protection

Good analytics does not need to know who a user is by name. You need patterns: how many people, from which channel, at which step they dropped off. That principle, data minimization, protects you and your customers and fits the spirit of Saudi Arabia's Personal Data Protection Law, overseen by SDAIA, and the UAE's federal data protection law.

• Keep names, phone numbers, emails and addresses out of event properties.
• Explain in your privacy policy what usage data you collect, why, and which outside tools you use.
• Ask for consent to tracking and advertising cookies where required, and actually respect a refusal.
• Delete old event data you no longer need after a period you define.
• If your sector requires data to stay in the country, check that your analytics tool does not move it abroad without a basis.

From numbers to decisions: a simple weekly routine

1. Every Sunday morning, open the same dashboard and look at the six numbers against last week and the same season.
2. Find the biggest leak in the journey: where do most people stop?
3. Pick one change for that leak, such as a simpler checkout, adding Apple Pay or clearer delivery fees.
4. Ship the change and note the date in a simple log.
5. Review the result after two weeks before deciding to keep it or roll it back.

A/B testing is useful, but it needs enough visitors for a reliable result. If your store gets only a few orders a day, one considered change watched for two weeks is more honest than a test called after two days. To turn numbers into real growth, read from MVP to product-market fit and abandoned cart recovery.

Frequently asked questions

Terms in this article