Customer analyticsDevice financeAfrica

M-KOPA’s trust breaks at the finish line

An analysis of 21,832 Play Store reviews shows that the customers most likely to be disappointed are the ones who finished paying, and that this problem is growing in 2026.

21,832Google Play reviews analysed
4.09★average rating, 4.25★ in 2026
15.9%of 2026 reviews complain about loan access
36%of all helpful votes sit on those reviews

At a glance

  1. Loan access after repayment is the biggest problem, and it is growing. Cash-loan access and limits appear in 11% of all reviews and 16% of 2026 reviews. About half of these complaints come from customers who have finished paying, and 60% of all questions in reviews are about it.
  2. Star ratings hide it. 40% of loan-access complaints sit inside 4–5★ reviews, and these reviews hold 36% of all helpful votes (44% in 2026). The most-voted review is a five-star review complaining about it.
  3. Customers are not leaving for rivals. Of 441 leave signals, 368 are requests to remove the app after paying and only 1 names a rival. The threat is dormancy, not defection.
  4. Customers love the app, but they recommend the credit. Replies are fast (median 3 hours in 2026) but 55% use one template and 96% push to email. An eligibility status screen is the first pilot.

The headlineM-KOPA customers love the phone and the app; the trouble starts when they finish paying

Across 21,832 Play Store reviews, M-KOPA averages 4.09 out of 5 (4.25 in 2026), and 69% of reviewers give five stars. Customers praise the app’s ease of use, the phone and the access to credit. On the surface, this is a healthy product.

The review text tells a sharper story. One problem towers over the rest: customers who have finished paying for their phone and cannot get a cash loan. Loan access and limits appear in 11% of all reviews and 16% of 2026 reviews. About half of those complaints (48%; 52% in 2026) mention having completed repayment. These are customers who have, by definition, repaid in full, and they feel the promise of the next loan has been withdrawn.

This article uses the reviews to answer the questions a product and CX team would ask: why customers threaten to leave and where they go, how they compare M-KOPA with others, what they love and recommend, which features they ask for, whether updates have worked, and whether fixes would be worth the cost.

Exhibit 1

Cash-loan access, the largest problem in the full history, is even larger in 2026, so recent reviews should lead

Headline measures, all reviews versus 2026 (1 Jan to 3 Sep 2026)

MeasureAll reviews2026ChangeWhich window to trust
Average rating4.094.25+4%All-time is representative
Reviews rated 1–2★19.9%16.4%-18%Either; 2026 is moving
Reviews with a complaint24.5%23.4%-4%All-time is representative
Leave signals per 1,000 reviews20.216.2-20%Either; 2026 is moving
Top complaint: cash loan access and limits11.4% of reviews15.9% of reviews+39%Read 2026. Materially different
Top praise: ease of use and app experience33.5% of praise30.8% of praise-8%Either; 2026 is moving
Public replies by the company85.6%99.5%+16%Either; 2026 is moving

Note: A change of 20% or more (or 8% to 20% with a clear trend) is treated as material.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

2026 lens

Read 2026 for the main problem. Loan access is 15.9% of 2026 reviews, against 11.4% all-time, and customer-support and app-login complaints are both lower. The all-time view understates how concentrated the pain has become.

Leaving or complaining?Neither the complaint rate nor the leave rate is rising; one complaint is, and it is growing fast

The share of reviews carrying a complaint has stayed in a narrow band, between 22% and 29%, for nearly three years. Average rating rose from 3.86 to 4.31. Explicit leave signals have fallen: 20.2 per 1,000 reviews overall, 16.2 in 2026, with 88 in the first eight months of the year.

So “are customers actually leaving, or simply complaining more?” has a precise answer here: neither. What is changing is the mix. Reviews complaining about loan access rose from 65 in Q4 ’23 to 324 in Q2 ’26, an increase of 5.0x, while total reviews grew 2.7x. Loan access is taking a larger slice of a complaint rate that has not moved: complaint reviews grew with review volume, from 211 in Q4 ’23 to 482 in Q2 ’26, but loan access grew faster.

Exhibit 2

Complaint rates are steady and leave signals are falling, so customers are neither leaving more nor complaining more, but one problem is growing inside the total

Quarterly view of M-KOPA reviews, Q4 ’23 to Q3 ’26 (latest quarter partial)

A. Share of reviews, %

B. Reviews per quarter by rating, count

C. Reviews with a complaint, count

D. Reviews with a leave signal, count

Note: Q3 ’26 covers 1 Jul to 3 Sep only. “Complaint” means at least one negative comment about a specific aspect.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table

A. Share of reviews, %

Reviews with a complaintReviews rated 1–2★
Q4 ’2328.6%26.2%
Q1 ’2421.5%19.1%
Q2 ’2425.4%22.0%
Q3 ’2427.3%18.7%
Q4 ’2423.9%20.2%
Q1 ’2521.9%18.6%
Q2 ’2524.2%17.2%
Q3 ’2523.7%17.0%
Q4 ’2523.8%18.4%
Q1 ’2622.6%17.7%
Q2 ’2623.9%16.8%
Q3 ’2623.5%14.5%

B. Reviews per quarter by rating, count

Rated 4–5★Rated 1–2★
Q4 ’23516193
Q1 ’24779194
Q2 ’24676203
Q3 ’24829207
Q4 ’24838223
Q1 ’251,074260
Q2 ’251,632355
Q3 ’251,490322
Q4 ’251,252297
Q1 ’261,365306
Q2 ’261,597339
Q3 ’261,356242

C. Reviews with a complaint, count

Reviews with a complaint
Q4 ’23211
Q1 ’24219
Q2 ’24234
Q3 ’24302
Q4 ’24264
Q1 ’25307
Q2 ’25500
Q3 ’25449
Q4 ’25384
Q1 ’26391
Q2 ’26482
Q3 ’26393

D. Reviews with a leave signal, count

Reviews with a leave signal
Q4 ’2321
Q1 ’2415
Q2 ’2414
Q3 ’2418
Q4 ’2438
Q1 ’2534
Q2 ’2547
Q3 ’2555
Q4 ’2543
Q1 ’2628
Q2 ’2627
Q3 ’2633
So what

The overall satisfaction numbers are not the story. The risk is a single, rising complaint that sits among the company’s best customers, which is invisible in an average.

2026 lens

Use 2026 for any statement about direction. The all-time average mixes three years in which customer-support and app-login problems were larger and loan access was smaller.

What keeps coming backLoan access is the dominant recurring complaint and a loyalty problem the star rating cannot see

Cash-loan access and limits are the top complaint in 12 of 12 quarters. App login, updates and performance (5.7%) are second, followed by phone lock and unlock (2.4%) and customer support (2.4%). In 2026 the gap has widened: loan access is 15.9% of reviews and nothing else exceeds 3.4%.

Exhibit 32026 view recommended: the pattern has shifted

Loan access is twice as large as any other complaint, and in 2026 it is pulling further ahead

Share of reviews with a complaint in each theme, %

Note: A review can appear in more than one theme. Themes group aspect labels assigned by the tagging model (coverage of negative aspects: 91%).

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Share of all reviews (all)Share of all reviews (2026)
Cash loan access and limits11.4%15.9%
App login, updates and performance5.7%3.4%
Phone lock and unlock2.4%1.3%
Customer support2.4%1.7%
Phone and device quality2.0%1.3%
Payments and balance display1.9%1.0%
App removal and uninstall1.3%0.7%
Trust and honesty1.1%0.7%

The recurrence is only part of the problem. The reviews that raise loan access are the most visible and the most invisible at once. They carry 36% of all helpful votes on the platform (44% in 2026). And 40% of them are rated four or five stars, so a team watching the star rating never sees them. The most-voted review in the dataset, with 5,153 votes, is a five-star review that praises the app and then notes that, after paying off the phone, the customer was told they are no longer eligible. The second-most-voted (5,025 votes) is also five stars and also about a completed customer who cannot get a cash loan.

Exhibit 52026 view recommended: the pattern has shifted

Loan-access complaints are a loyalty problem: half come from people who finished paying, and four in ten are hidden inside 4–5★ reviews

Characteristics of reviews complaining about cash-loan access, %

Note: Based on 2,498 reviews with a loan-access complaint (862 in 2026). “Finished repaying” is detected from words such as paid off, completed, cleared, settled.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Loan-access complaints (all)Loan-access complaints (2026)
Mention having finished repaying48%52%
Rated 4–5★ despite the complaint40%42%
Share of all helpful votes held by these reviews36%44%
Exhibit 4

Loan access and app performance have been the two biggest problems in all 12 quarters; only loan access is growing

Share of reviews with a complaint, by theme and quarter, %

Note: Darker cells mean a larger share of that quarter’s reviews.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Q4 ’23Q1 ’24Q2 ’24Q3 ’24Q4 ’24Q1 ’25Q2 ’25Q3 ’25Q4 ’25Q1 ’26Q2 ’26Q3 ’26
Cash loan access and limits8.88.714.019.010.98.814.413.714.414.616.117.1
App login, updates and performance9.16.03.64.04.35.33.44.43.02.93.83.4
Customer support5.63.03.42.42.32.42.21.62.02.31.61.3
Phone and device quality1.81.63.61.92.63.11.92.11.81.91.01.0
Phone lock and unlock3.31.72.11.02.83.42.11.41.82.01.30.5
App removal and uninstall1.60.71.00.52.52.31.71.71.70.60.70.8
Payments and balance display2.31.81.21.20.91.41.11.21.20.81.10.9
2026 lens

Read 2026. Customer support (from 3.5% to 1.8% of reviews) and app login (from 5.7% to 3.3%) have improved; loan access has risen from 12.6% to 15.5%.

Why customers leaveCustomers are not leaving for a rival; they are leaving because the loan has ended

M-KOPA has 441 reviews with a leave signal, 20.2 per 1,000 reviews. The surprise is what they are. 368 (83%) are uninstall requests, and about half of those mention having finished paying; the customer wants the app and its controls removed. Only 72 say they will stop using the service and just 1 is flagged as switching to a rival. Across all 441 leave-signal reviews only five mention any competitor at all (Sun King three times, Watu twice).

Exhibit 6

Leave signals are mostly requests to remove the app after paying, not defections to a rival

Reviews with a leave signal, by type, count

Note: 441 reviews in total. Of the 368 uninstall requests, 51% mention having finished paying; their average rating is 2.0★.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Reviews
Asked to uninstall the app368
Said they will stop using M-KOPA72
Named a rival they are moving to1

Behind the explicit signals, the reasons again point to the end of the loan. After app removal, loan access is the leading theme in leave-signal reviews (20%), followed by phone lock and unlock (13%), app performance (13%) and phone and device quality (which carries a leave lift of 3.9x). Phone lock and unlock has a leave lift of 4.0x, so it is a smaller problem that is more likely to end the relationship.

Exhibit 72026 view recommended: the pattern has shifted

App-removal requests dominate leave signals, followed by loan access and phone lock

Share of reviews mentioning each theme, among all complaining reviews and among reviews with a leave signal, %

Note: Leave signals: 441 overall, 88 in 2026. App removal appears at 70% in leave signals because uninstall requests are counted as leave signals. Shares are of the 321 leave-signal reviews that carry a tagged complaint (the priority table later uses all leave signals).

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Among all complaining reviewsAmong reviews with a leave signalAmong all complaining reviews (2026)Among reviews with a leave signal (2026)
App removal and uninstall5.3%70.1%3.1%58.3%
Cash loan access and limits46.6%19.9%68.1%38.3%
Phone lock and unlock9.8%13.1%5.5%6.7%
App login, updates and performance23.2%12.8%14.4%5.0%
Phone and device quality8.2%10.9%5.5%11.7%
Customer support9.7%9.0%7.4%3.3%
So what

M-KOPA does not lose customers to competitors in reviews; it fails to give completed customers a reason to stay. The threat is dormancy, not defection, and the response is a better post-payoff offer, not a retention discount.

How customers compareM-KOPA is rarely compared with anyone, and when it is, the comparison is about credit

Customers rarely benchmark M-KOPA: 150 reviews in 21,832 (0.7%), 28 of them in 2026. That is a sign of a product with few visible substitutes at the point of purchase. The brands named are mobile-money lenders (Safaricom products such as Fuliza, M-Pesa and M-Shwari) and other pay-as-you-go and instalment providers (Watu, Sun King).

The comparisons that exist focus on loans. In comparison reviews overall, 65% are rated 4–5★ and 32% 1–2★. The leading criticism is cash-loan access (18 reviews) and the leading praise ease of use and loans and financing access. When Sun King is the comparator, 67% of reviews are rated one or two stars, although the group is small.

Exhibit 8

Comparisons are rare, and the few that name a rival centre on loans

Ratings given in reviews that compare M-KOPA with a named rival, %

Note: Only 150 reviews (0.7% of the total) compare M-KOPA with another provider; groups with fewer than 10 reviews are indicative only.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Review rated 4–5★Review rated 1–2★
Watu (n = 7)43%43%
Sun King (n = 6)17%67%
M-Pesa / Safaricom (Fuliza, M-Shwari) (n = 6)67%33%
2026 lens

There is not enough 2026 comparison data to separate from the all-time view. Use the all-time picture and treat rival names as anecdotes, not market share.

What they loveCustomers love the app, but they recommend the credit

The most praised aspects are ease of use and the app experience (33% of reviews that praise something specific), the phone itself (27%) and access to loans and financing (19%, rising to 22% in 2026). Trust and reliability climbed from 13% to 20% in 2026.

Exhibit 92026 view recommended: the pattern has shifted

Customers love the app and the phone; ease of use is the most-praised aspect

Share of reviews praising a specific aspect that mention each theme, %

Note: Generic praise is excluded.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Share of reviews that praise something specific (all)Share of reviews that praise something specific (2026)
Ease of use and app experience33.5%30.8%
Phone and product quality27.0%24.6%
Loans and financing access19.3%21.6%
Trust and reliability12.9%20.0%
Customer support12.1%10.1%
Payment flexibility and transparency8.6%7.3%

Recommendations are different. Among the 193 reviews that recommend M-KOPA for a specific reason, loans and financing access leads at 36%, 1.9 times as common as in other praise, followed by the phone (34%) and payment flexibility (2.0x). Ease of use, the most-praised aspect, is under-represented among recommenders (0.8x). The most common phrases in recommendations are “cash loan”, “best loan” and “recommend everyone”.

The same feature is therefore the main reason to recommend and the main reason to complain. Credit access is what makes customers advocates when it works, and what costs the loudest reviews when it does not.

Exhibit 102026 sample is thin: read with caution

But what makes them recommend M-KOPA is access to credit and flexible payment

Share of recommending reviews with a specific reason that mention each theme, versus all praising reviews, %

Note: 375 reviews recommend M-KOPA; 46% give no specific reason, and 193 do. The 2026 sample of recommenders with a reason is small.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Reviews that recommend (all)Reviews that recommend (2026)All praising reviews (all)All praising reviews (2026)
Loans and financing access36%34%19%22%
Phone and product quality34%26%27%25%
Ease of use and app experience28%28%34%31%
Payment flexibility and transparency17%16%9%7%
Customer support15%18%12%10%
Trust and reliability15%26%13%20%
So what

Lead acquisition messaging with the phone and the app, and lead referral messaging with credit. Then fix the post-payoff experience, because the product’s best advocacy asset is also its biggest source of public disappointment.

What they ask forCustomers ask for more credit first, and for control over the phone lock second by votes

Of 500 reviews with a feature request, the largest group wants higher loan limits and easier cash-loan access (155 reviews, 9,111 votes), followed by app updates, login and notifications (83), payment options and balance display (62), phone-lock flexibility such as grace periods and temporary unlock (51 reviews, 7,255 votes), and offline and low-data use (49 reviews, 6,841 votes). A wider range of phones and TVs gets few requests but strong endorsement (4,842 votes).

Exhibit 112026 view recommended: the pattern has shifted

Higher loan limits dominate requests; phone-lock flexibility and offline use draw the most votes per request

Reviews asking for each kind of feature, count (votes in the tooltip and data table)

Note: Based on 500 reviews with a request (116 in 2026). Votes count helpful votes on the whole review. “Other requests” left out.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Reviews asking for it (all)Reviews asking for it (2026)
Higher loan limits and cash loan access15553
App updates, login and notifications8324
Payment options and balance display6210
Phone-lock flexibility (calls, grace period, temporary unlock)5111
Offline and low-data use492
Easier app removal after payoff263
Wider product range (phones, TVs)233
Rewards for loyal customers155

The chart shows how many customers ask for each feature; the table below shows what they describe. The loan requests are specific: 34 ask for a bigger limit (some name an amount), 19 ask for loans to continue after the phone is paid off, and 9 argue that on-time payers deserve more. The phone-lock requests are a short list of small permissions rather than a call to remove the lock: calls and contacts, Wi-Fi or data access, a longer temporary unlock and a visible days-left count.

Exhibit 12

Behind the counts, requests are small permissions: a bigger loan limit after payoff, calls and Wi-Fi on a locked phone, and a way to pay without data

What customers describe, by topic

TopicReviewsWhat customers describe
Higher loan limits and cash loan access15534 ask for a bigger limit, and several name an amount (for example KSh 28,000 or 50,000). 19 ask for loans to continue after the phone is fully paid off, 9 argue that customers who pay on time deserve more, and 13 want a loan offered quickly or automatically once the last one is cleared. One asks to see the limit before deciding whether to pay a loan off early.
App updates, login and notifications83Almost all are a plain call to update or fix the app (“update M-KOPA”, “need the latest update”). 10 ask for payment notifications, such as the app updating itself with an alert after a payment, and 5 for login or OTP fixes. A few ask for agent tools, such as a view of which customers pay slowly so they can be reminded.
Payment options and balance display62About 20 ask to see the remaining balance, payment schedule and exact end date inside the app. Others ask for more ways to pay, such as linking a bank account, and for one paybill and app that agree with each other, and for payments to be deducted from the outstanding balance clearly.
Phone-lock flexibility (calls, grace period, temporary unlock)5119 ask for Wi-Fi, data or phone-settings access while locked, including buying data bundles or airtime when locked. About 10 ask for calls and contacts to keep working while locked, with some adding messages and music. 6 ask for a longer temporary unlock or grace period, 5 for the days left to be shown rather than a lock date, and some for automatic unlock as soon as a paybill payment lands.
Offline and low-data use4930 ask for the app to work without data or offline, mostly to pay or unlock without buying bundles (“it becomes costly”). About 20 ask for free or bundled data for the app (one asks for 2 GB a month) and 5 for a USSD or SMS alternative.
Easier app removal after payoff26These reviews ask to uninstall M-KOPA or lift its lock once payment is complete. About 9 name restrictions to release, such as build-number and developer-settings access, tracking and the lock features.
Wider product range (phones, TVs)23About 12 ask for products beyond phones, such as TVs, laptops, fridges, solar and batteries; 10 ask for other phone brands (one says some customers fear Samsung screen repairs) and 6 ask to choose their own offer instead of picking from a short list.
Rewards for loyal customers15Customers who have finished paying ask to be thanked or rewarded (3 mention a prize or gift), and 11 link rewards to priority access to cash loans.

Note: Review counts are those in Exhibit 11. Sub-request counts come from keyword rules applied to the text of each request, were spot-checked by reading samples, and can overlap or undercount (requests written in other languages are the main gap). Requests are paraphrased from the reviews.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

Two of these are cheap to test without changing credit policy: showing the remaining balance, schedule and end date in the app, and an offline or USSD way to confirm a payment. Both answer requests customers make in plain terms and sit close to the payment and balance complaints elsewhere in this analysis.

Which feature to try first? Loan access scores highest on evidence, but it is a credit policy decision. The first pilot should be one that solves the problem without changing underwriting: an in-app eligibility status screen that tells a customer whether they qualify for a cash loan, why not if they do not, and what happens next. Over 60% of all questions in reviews (1,302 of 2,160) are this one question. Easier app removal after payoff is the cheapest second pilot.

Exhibit 13

Loan access scores highest but is a policy change; removal and balance fixes are the cheap pilots

Pilot priority for six requested features

Feature to pilotProblem solvedAsksVotes2026 asksLeave shareEffortScore
Higher loan limits and cash loan accessLoan access after repayment1559,1115315%High (credit policy)71
Easier app removal after payoffUninstall requests after payoff26146351%Low46
Phone-lock flexibility (calls, grace period, temporary unlock)Phone lock and unlock517,2551110%Medium39
Offline and low-data useApp access with little data496,84129%Medium37
App updates, login and notificationsLogin and update friction832,566249%Medium33
Payment options and balance displayPayments not reflecting62899102%Low18

Note: Asks = reviews requesting it; votes = helpful votes on those reviews; leave share = share of leave signals tied to the problem the feature solves. Score = 35% asks + 25% votes + 40% leave share, scaled to the top item. Effort is a judgement, not a measurement.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

2026 lens

Loan access is an even larger share of requests in 2026 (53 of 116, 46%, against 31% all-time), so the case for it strengthens. Use 2026 to choose and all-time to estimate demand.

How customers behaveA complaining M-KOPA customer asks for help, reports a bug and warns others in one review

Among complaining reviews, 13% also ask for help (lift 1.9), 12% report a bug (lift 2.8) and 8% warn other customers (lift 3.9, the strongest of the three). That is much more than at Moniepoint, where 2.4% do. A warning to others is a public cost: it appears next to the install button.

About 24% of complaining reviews also contain praise, so a quarter of complainers are still fans. Complaint reviews are also more likely than average to be questions: 369 complaining reviews also ask for information (lift 2.0). That makes them addressable by a clear answer rather than a concession.

Exhibit 142026 view recommended: the pattern has shifted

Complaining M-KOPA customers are far more likely to warn others than Moniepoint customers

Share of complaining reviews that also do each of the following, %

Note: Based on 5,521 complaining reviews.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table
Share of complaining reviews that… (all)Share of complaining reviews that… (2026)
Also praise something18.6%23.3%
Also ask for help13.1%16.9%
Also report a bug11.7%10.2%
Also warn other customers8.3%6.4%
Also suggest a feature5.0%4.0%
Also say they will leave1.5%2.1%
Also compare with a rival1.0%0.6%
Also ask for a refund0.3%0.2%
Exhibit 15

Complaints about loan access travel with support and trust concerns

Complaint themes that appear together, ranked by lift

Complaint AComplaint BReviews with bothLiftWhen A appears, B also appears
Customer supportTrust and honesty411.7x8%
Customer supportPhone and device quality591.3x11%
Customer supportPhone lock and unlock621.2x12%
Cash loan access and limitsTrust and honesty1010.9x4%
Phone and device qualityPhone lock and unlock400.9x9%
App login, updates and performancePayments and balance display830.8x7%

Note: Lift above 1 means the two themes appear together more often than chance would predict.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

Do updates work?Updates have worked in aggregate, with app complaints down from 8.1% to 3.3% of reviews, but no single release can claim it

For early versions, 8.1% of reviews carry an app login, update or performance complaint; for recent versions the figure is 3.3% (a difference beyond chance, Spearman -0.55). The theme is also lower in 2026 (3.4%) than in the full history (5.7%). In absolute terms the number of reviews with this complaint has been roughly flat (between 33 and 84 a quarter) while total reviews roughly doubled, so some of the fall is dilution.

Individual releases are hard to credit: of 66 version-to-version comparisons, 4 show a significant improvement, 3 a regression and 59 no difference, about what chance would produce. The honest conclusion is that app quality improved gradually and modestly, and that the one complaint that updates cannot fix, loan access, is a policy issue, not a release issue.

Exhibit 16

App-login and performance complaints fell by almost two-thirds per review, but individual releases cannot be credited

Share of reviews with an app login, update or performance complaint

A. By app version, oldest to newest

B. By month

Note: Versions with at least 120 reviews (67 of them), in order of first appearance. Early versions 44.9.0 to 2024.196.2; recent versions 2025.209.5 to 2026.176.4.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

View data table

A. By app version, oldest to newest

Share of reviews on that version with the complaint
44.9.017.4%
50.1.05.5%
52.0.09.5%
53.0.319.8%
55.3.39.5%
57.1.39.3%
61.2.312.9%
64.3.19.7%
78.2.115.3%
2023.72.312.1%
2023.195.311.1%
2023.207.212.2%
2023.259.15.6%
2023.270.45.3%
2024.23.36.3%
2024.65.02.0%
2024.99.72.9%
2024.132.33.4%
2024.162.33.9%
2024.174.24.5%
2024.184.21.9%
2024.196.26.8%
2024.206.13.2%
2024.213.25.8%
2024.224.42.0%
2024.231.53.2%
2024.252.49.5%
2024.252.54.8%
2025.13.25.0%
2025.24.27.5%
2025.33.23.1%
2025.44.45.2%
2025.60.43.0%
2025.69.54.5%
2025.80.12.9%
2025.91.32.2%
2025.99.12.6%
2025.107.53.7%
2025.117.45.0%
2025.142.15.1%
2025.156.24.0%
2025.165.22.5%
2025.184.32.3%
2025.191.14.2%
2025.201.22.7%
2025.209.51.6%
2025.223.52.6%
2025.232.33.0%
2025.254.32.2%
2026.5.31.2%
2026.16.32.7%
2026.26.52.3%
2026.36.94.4%
2026.46.43.2%
2026.57.42.6%
2026.69.44.1%
2026.80.22.6%
2026.89.44.8%
2026.100.23.3%
2026.111.63.5%
2026.120.55.8%
2026.133.43.8%
2026.143.33.9%
2026.156.33.8%
2026.156.44.2%
2026.166.62.0%
2026.176.43.8%

B. By month

Share of that month’s reviews
2024-043.2%
2024-053.6%
2024-063.8%
2024-074.7%
2024-083.7%
2024-093.5%
2024-105.0%
2024-112.2%
2024-125.7%
2025-015.6%
2025-026.2%
2025-034.7%
2025-043.8%
2025-053.5%
2025-063.0%
2025-075.2%
2025-084.7%
2025-093.4%
2025-103.4%
2025-112.8%
2025-122.8%
2026-012.1%
2026-023.2%
2026-033.1%
2026-043.1%
2026-053.4%
2026-064.8%
2026-073.4%
2026-083.3%
2026-093.5%
2026 lens

The recent-version view is the right one for app questions. The gains were mostly made between 2024 and 2025. For loan access, no version has changed the picture.

Reputation and brand voiceM-KOPA is fast and consistent in public, but its voice does not match the problem customers describe

The response record is strong: 99.5% of 2026 reviews have a public reply, with a median time of 3 hours (75% of replies within a day overall). Replies are respectful. They are also uniform: 55% of all replies use one template that begins by saying the company is glad the customer is enjoying their M-KOPA experience, and 96% direct the customer to email, often with a ticket reference (60%).

The mismatch appears where the pain is. Among replies to loan-access complaints, 10% use the “glad you’re enjoying” opening. For app-login complaints it is 8%. Moving a conversation to email is sensible for account-specific problems; but a public question that thousands of customers share (“why can’t I borrow after finishing?”) is better answered once, publicly, in the reply itself.

The reputation exposure is concentrated. The top 1% of reviews hold 82% of helpful votes. In the 20 most-helpful reviews of all time, 15 contain a complaint; among 2026 reviews, 14 of the top 20 are rated one or two stars and 18 contain a complaint. A new visitor who sorts by most helpful meets the loan-access problem first.

Exhibit 17

M-KOPA replies to almost every review within hours, but with one template and a pointer to email

Reputation and reply indicators

IndicatorAll reviews2026
Public reply rate85.6%99.5%
Median time to reply (hours)8.53.1
Replies inside 24 hours75%n/a
Replies that send the customer to email or another channel96%n/a
Replies that ask for a rating2%n/a
Replies that apologise24%n/a
Share of replies that use the single most common template55%n/a
Replies to 1–2★ reviews that use celebratory wording5%6.2%
Share of all helpful votes sitting on 1–2★ reviews45%n/a
Share of helpful votes held by the top 1% of reviews82%n/a
Of the 20 most-helpful reviews, rated 1–2★814
Of the 20 most-helpful reviews, containing a complaint1518
5★ reviews that still contain a complaint8.9%9.4%

Note: Ratings can be edited after a reply, which affects reply-quality indicators. Helpful votes show what visitors see when the store sorts by “most helpful”.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

So what

The audit of M-KOPA’s public voice finds speed, consistency and a gap in substance. A tiered reply library, with the eligibility explanation written into the public reply, would convert the best-read negative reviews into evidence of service.

Content ideasThe reviews already contain the FAQ: one question is asked 1,302 times

10% of reviews contain a question or help request, and 60% of those are about loan access after finishing payment. That one question was asked 416 times in 2026. It should be answered in four places: the in-app status screen, a pinned FAQ, the reply template and a short explainer video.

Exhibit 18

One question accounts for 60% of all questions in reviews, and it is a content gap as much as a product gap

Questions asked in reviews (help requests and information seeking), and the asset each calls for

Question customers askReviews askingAsked in 2026Asset to build
Why can I not get a cash loan after finishing payment?1,302416Eligibility explainer, an in-app status screen and a pinned public reply
When will my phone unlock after I pay?17631Unlock-timing explainer with an example timeline
How do I uninstall the app after payoff?17526Short “after you finish” guide and one-tap close-out
How do I update the app or log in?15727Login and update help article
Why has my payment not reflected?12337Payment-status explainer and confirmation message
How do I reach support?4911One visible contact path repeated in every reply

Note: 2,160 reviews (9.9% of all) contain a question or help request. Themes assigned by keyword; 167 fit no theme.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

The proof points are in recommenders’ own language: “cash loan”, “best loan”, “recommend everyone”, “customer care”. These are customer-validated claims for referral prompts and campaigns. “Customer care” appears nine times, which is a small but usable signal for a content series on resolved cases.

Exhibit 19

The words recommenders use are ready-made proof points

Most common two-word phrases in reviews that recommend M-KOPA

Phrase customers useReviews
cash loan24
best ever18
recommend everyone13
best loan10
cash loans9
customer care9
recommend anyone9
best company8

Note: Phrases are counted once per review; brand-name variants removed.

Source: Google Play Store reviews of the M-KOPA app (n = 21,832; 2020 to Sep 2026); analysis by Adediran Adeyemi

Is the fix worth it?For M-KOPA, the case for fixing loan access rests on repeat borrowing, not on preventing churn

Loan access is the largest problem by volume (15.9% of 2026 reviews) but the weakest churn driver of the main themes: reviews that mention it carry leave signals at only 1.3 times the average rate. A churn-based calculation would therefore understate the value of a fix. The customers writing these reviews do not threaten to leave; they ask to borrow again.

The value at stake is the next loan to customers who have already shown they repay. Roughly half of loan-access complaints come from people who have finished paying (52% in 2026). If even a modest share of them would take a repeat loan, the revenue is immediate and comes from customers with the lowest acquisition cost. In contrast, phone lock and unlock and device quality are smaller themes with leave lifts of 4.0x and 3.9x, the opposite profile: few customers, high churn risk.

Will the fix pay for itself? Try your own numbers

Pick a complaint theme. For M-KOPA, treat “who would leave” as the share of affected customers who would otherwise have taken a further cash loan. The default is the review-based excess leave rate, which is low for loan access; replace it with your own estimate.

0customers affected
0customers at risk of leaving
0annual revenue protected
0return on cost, year one
0customers to keep to break even

Reviews come from customers who chose to write. Treat the affected and leave rates as indicators, not census figures. Revenue per customer and the effect of the fix are placeholders until you replace them.

Read with care. Reviews cannot show credit risk. A cash-loan fix needs underwriting data on whether post-payoff customers are creditworthy. The review evidence shows demand and disappointment, not repayment capacity.

Where to startFive moves for the next quarter, in order

The analysis points to five actions, ordered by evidence strength and by the cost of being wrong.

  1. Ship an eligibility status screen

    Loan-access complaints are the largest and fastest-growing theme, and 60% of all questions in reviews are about it. Telling customers where they stand, and why, answers the question without changing underwriting.

    Owner
    Credit product and design
    Measure
    Loan-access share of reviews; share of questions; helpful votes on loan-access reviews
    First test
    Show the status screen to 20% of customers who have completed payment; compare loan-access complaints and repeat-loan applications after four weeks.
  2. Write a post-payoff journey

    Half of loan-access complaints come from customers who finished paying; app-removal requests have the same origin. Treat completion as a moment, not an ending.

    Owner
    Customer lifecycle
    Measure
    Share of completed customers who take a further product; uninstall requests; ratings after completion
    First test
    Send a completion message that explains eligibility, offers app removal and states the next steps.
  3. Retire the “glad you are enjoying” template on complaints

    One template covers 55% of replies, including replies to loan-access and app complaints. Public replies that ignore the problem cost credibility.

    Owner
    Customer experience
    Measure
    Replies to 1–3★ reviews that address the specific issue; follow-up rating changes
    First test
    Write one public reply per top-five complaint theme and pin them.
  4. Count the hidden complaints

    9% of five-star reviews and 40% of loan-access complaints rated 4–5★ show that stars hide the main problem. Tag reviews by theme, not just by rating.

    Owner
    CX analytics
    Measure
    Complaint reviews per 10,000 active customers; share of 4–5★ reviews with complaints
    First test
    Add the theme view to the monthly CX report.
  5. Test phone-lock flexibility

    Phone lock and unlock is a smaller theme with a 4.0x leave lift and 7,255 votes on requests for flexibility (grace periods, temporary unlock).

    Owner
    Product and credit risk
    Measure
    Phone-lock complaints; leave signals; payment timeliness
    First test
    Trial a short grace window for customers with a clean record in one region; check repayment behaviour before extending.

Data and methodAbout this analysis

What we analysed

21,832 public Google Play Store reviews of the M-KOPA app, covering 2020 to 3 September 2026 (most from 2023 onward). Each review was split into segments and tagged with aspects, sentiment per aspect, customer intents, churn signals and competitor mentions.

How the measures are defined

  • Complaint: a review with at least one negative segment about a specific aspect.
  • Leave signal: an explicit statement of leaving, switching or asking to uninstall. Uninstall requests are included and shown separately.
  • Theme: a group of aspect labels assigned by transparent keyword rules; 91% of negative and 87% of positive aspects map to a theme.
  • 2026 view: reviews dated 1 January 2026 onward (5,418 reviews).
  • Finished repaying: detected from words such as paid off, completed, cleared, settled in the review text.

Limits to keep in mind

  • Reviewers choose to write and are not a random sample of customers.
  • Tagging was not checked review by review.
  • The dataset has no country field, so markets cannot be separated.
  • Ratings can be edited after a reply is posted.
  • Helpful votes are on whole reviews.
  • Reviews cannot show revenue, credit risk or true churn.

Have thousands of reviews nobody has properly read?

I turn app-store reviews, surveys and support logs into counts, causes and a plan your product team can act on. The first conversation is free.