How to Monitor Bank App Reviews and Customer Feedback

A bank releases an app update on Sunday night. By Monday lunchtime, customers have reviewed it. Some leave stars in the app store. Many more reply under the bank's latest post on X, comment on a TikTok that explains the new screen, or mention it in a Google Maps review of their branch.

Few of them tag the bank. Most write in Arabic, in Saudi, Emirati or Egyptian dialect, and switch to English mid-sentence. The quarterly survey will ask about the same update weeks later, after customers have already answered in public.

This playbook shows how to monitor customer feedback for banks in Arabic: where to look, how to sort it, who acts on it, and how to read a release against two rival banks. It ends with a one-page tracker and a worked example.

Step 1: Know where bank customers give feedback

Step 2: Sort feedback with a retail banking topic map

Sort every mention into topics your teams own. One overall sentiment score hides more than it shows. Here is a starter map in the region's own phrasing. The phrases are illustrative.

Two habits make this harder than it looks:

• Code-switching. "الـ app بعد الـ update صار يعلّق عند الـ login" (Saudi: since the update the app freezes at login) puts the topic in English words inside an Arabic sentence.
• Untagged and misspelled mentions. Customers write the app's nickname, the card's name, or the bank's name spelled three ways, in Arabic and Latin letters.

Step 3: Read sentiment and emotion separately

Sentiment tells you whether a topic is praised or criticised. Emotion tells you what sits behind it. The same negative score can call for very different fixes:

• Negative with frustration, right after a release: a usability problem. Digital fixes the flow, fast.
• Negative with disappointment on fees or offers: an expectations gap. Explain the terms better, or review them.
• Positive with relief on onboarding: a feature that removed effort. Promote it in customers' own words.
• Mixed with confusion on rewards: customers like the product but cannot use it. Fix the explanation, not the programme.

Step 4: Run a weekly customer feedback review

Forty-five minutes, same day every week. Sunday suits a Sunday to Thursday week. Keep the agenda fixed:

1. Read the week by topic (10 min): share of conversation, sentiment and leading emotion for each topic.
2. Check what changed (10 min): app releases, new offers, fee updates, campaigns, salary week.
3. Pick the three topics that moved most (5 min).
4. Give each one an owner, one action and a date (10 min).
5. Close last week's loop (10 min): did last week's fix change what customers say?

Step 5: Track every app release on one page

Use this app release feedback tracker for the two weeks after each update.

Fill in first: app version and release date · what changed · listening terms (app name and nicknames, misspellings, Franco-Arabic spellings, new feature names) · two rival banks · platforms, including Google Maps reviews and app-store reviews from your console · dialects (Saudi, Emirati, Egyptian, plus English and Franco-Arabic).

How to read it: a topic that grows with rising frustration is this week's fix. A topic that grows with relief or surprise is something to promote. If a rival's column moves the same way as yours, the cause is the market, not your release.

Step 6: Benchmark against two rival banks

1. Use the same topic map for all three banks, so you compare like with like.
2. Use the same weeks. Salary week, Ramadan offers and back-to-school spending change what everyone talks about.
3. Compare share, not volume. A bigger bank always has more mentions. Compare each topic's share of each bank's conversation.
4. Read sentiment and emotion side by side. Two apps can both score positive while one draws relief and the other mild disappointment.
5. Separate creators from customers. One explainer with a large audience can move a whole week.
6. Read each market on its own. Customers in Saudi Arabia, the UAE and Egypt raise different topics. Compare Saudi with Saudi.

Our competitor analysis guide covers the full method, and dima's market insights runs the comparison continuously.

Worked example: Bank A's app update, read over two weeks (illustrative)

Illustrative example. Bank A is a fictional regional digital bank with customers in Saudi Arabia, the UAE and Egypt. It releases an update with a redesigned transfer screen and a new rewards page. Bank B and Bank C, also fictional, are the rivals. Nothing below is real data.

1. Set-up. The tracker lists the app's name in Arabic and English, common misspellings, its nickname and Franco-Arabic spelling, and the words "التحويل" (transfer) and "النقاط" (points). Banks B and C are tracked the same way.
2. Days 1 to 3. Transfers become the largest topic in Bank A's app conversation. Sentiment is negative and the leading emotion is frustration: "فين زرار التحويل؟ قلّبت التطبيق كله" (Egyptian: where is the transfer button? I went through the whole app). Most of these posts do not tag the bank, and several misspell the app's name.
3. Days 4 to 7. A creator posts a TikTok explaining the rewards page. Comments are positive and the leading emotion is surprise: "ما كنت أدري إن النقاط تتحول كاش باك" (Emirati: I didn't know the points turn into cashback).
4. Week 2. Digital adds an in-app tip pointing to the new transfer button. Frustration on transfers falls and the leading emotion moves to satisfaction: "تعودت على الشكل الجديد، صار أسرع" (Saudi: I got used to the new look, it's quicker now).
5. Against rivals. Bank B's transfer conversation stays flat over the same weeks, so the jump came from Bank A's release, not the market. Bank C still leads on rewards, but Bank A closes part of the gap.
6. Decisions. Digital keeps the redesign and adds the tip to onboarding. Cards writes a short rewards FAQ in plain Saudi, Emirati and Egyptian phrasing. Marketing briefs two more creators on the rewards page. CX adds the transfer-button answer to call-centre scripts.

What actually decides it

A tracker is only as good as the mentions behind it. Bank customers rarely tag the bank. They write the app's nickname, the card's name or the branch, misspell all three, and switch between Arabic, English and Franco-Arabic in one line. A tool that tracks only the handle sees a thin slice of your customers, and a different slice of your rivals'. dima captures tagged, untagged and misspelled mentions with 100% coverage, and reads them in the dialect they were written in. That is the specific problem it was built to solve.

Choosing a platform for this work? See the best social listening tools for banks in MENA, dima's social listening and analytics, or our guide on protecting your bank's position in the digital market.

How much of what your customers said about your last app update ever reached your dashboard?

See what customers say about your next app release, including posts that never tag you.