Today, two weeks after launching My OneApp to considerable fanfare at the Decode 4.0 developer summit, Safaricom has issued a formal public apology. It appeared on their Twitter/X and Facebook pages this afternoon. It acknowledged poor experience. It called out roaming and diaspora customers specifically. It noted the particular unfairness of the situation for those who were automatically migrated to the new app through auto-update settings without any warning. It committed to fixing things. It was signed, simply, by Safaricom PLC.

It was, by most measures of a social media crisis communications, a reasonable apology. It was also, by most measures of crisis management timing, at least one week late. The reply threads it has generated on Twitter/X and Facebook is more analytically revealing and more commercially alarming than any social media listening report can tell. 

I have written about social media crises in Kenya for well over a decade. In 2013, I wrote about Art Caffe’s racism incident and the three steps any brand needed to take to manage that kind of crisis. In 2022, I documented Co-op Bank’s when a tweet about M-PESA deposit charges triggered a KOT firestorm that forced a complete policy reversal within hours. 

In 2024, I tracked how KFC Kenya’s chips crisis — triggered by the revelation that their chips were sourced from Egypt rather than Kenyan potatoes — eventually led to a genuine and measurable business transformation that I photographed on a billboard two years later. 

Each of those cases taught something specific and enduring. The OneApp crisis in many ways captures all of them, adds several new lessons that Kenya’s digital landscape has not previously had cause to generate, and then tops everything with a reply thread that captures, in the raw and unfiltered words of Safaricom’s own customers, exactly what a company that has spent 25 years building Kenya’s most trusted brand looks like when it loses the room completely.

Defining What We Are Actually Dealing With

A social media crisis is not simply a lot of people complaining online. It is an event that can have a material negative effect on a brand’s, company’s, or individual’s reputation — one that either originates offline and spreads to social media, or begins on social media and escalates beyond the brand’s capacity to contain it. 

The distinction that separates a genuine crisis from routine negative feedback is the simultaneous presence of six key drivers: volume shock, reach shock, sentiment inversion, User Generated Content (UGC) acceleration, persistence, and cross-channel spread. 

When all six are present together, you have a crisis. When all six are present in the context of a platform like M-PESA that 40 million people in Kenya depend on for daily financial transactions, the stakes are categorically higher than almost anything else a Kenyan brand has previously had to manage.

The OneApp situation clears every single one of those thresholds. This is not a matter of interpretation. It is measurable.

What The Social Media Data Shows

Using a dashboard tracking the keyword “OneApp” across the period of 30th March to 13th April 2026 against the preceding period of 15th to 29th March, the crisis signature is unambiguous across every relevant metric. 

Total mentions jumped from 43 to 292 — a 579 per cent increase. Social media mentions moved from 14 to 157, registered as greater than 1,000 per cent growth. Non-social mentions grew from 29 to 135, a 366 per cent increase. User-generated content surged from 23 to 196, a 752 per cent jump. Social media reach leaped from 820 to 245,000. 

Non-social reach moved from 381,000 to 1.0 million. The Average Presence Score moved from 15 to 22, a 47 per cent increase. The Advertising Value Equivalent jumped from $43,000 to $138,000–221 per cent more brand visibility, entirely of the wrong kind.

The sentiment picture is the most diagnostic element. Negative mentions rose from 3 per cent of the conversation to 26 per cent. Positive mentions declined from 14 per cent to 9 per cent. When neutral sentiment is stripped out entirely, the picture sharpens into something genuinely alarming: approximately 74 per cent negative versus 26 per cent positive in the current period, against approximately 86 per cent positive in the preceding period. 

This is not a drift toward negativity. It is sentiment inversion — a full reversal of the emotional polarity of the brand conversation — the classic and unmistakable signature of a crisis that has passed beyond the point of routine management.

The reach time-series shows a major spike around 8th April reaching approximately 430,000 at its peak, followed by sustained elevated reach across subsequent days. The weekly AI summary states the conversation reached 1.4 million users during the peak week, with a daily reach of 134,833 on 10th April. 

Critically, the mentions chart does not spike once and disappear. It ramps, peaks, and stays materially above the pre-crisis baseline through 13th April — the pattern that keeps a crisis alive long enough for mainstream media pickup and, more significantly, for it to become a reference point in how customers describe their relationship with the brand going forward.

The channel mix analysis shows the conversation spread across multiple source types —Twitter/ X, web, news, and other social platforms. Cross-channel spread is significant because it means the story is not containable by managing one platform. It has bridged from consumer social media into editorial coverage and general news — a progression that transforms what could have been a contained product complaint into a durable reputational narrative.

The Intelligence — & The Insight That Changes The Communications Strategy

Covering “OneApp” on Twitter/X between 12th and 13th April shows 198 posts from 153 contributors, generating 134,946 actual impressions, with total engagement comprising 532 likes, 227 reposts, 218 replies, 36 quotes, and 71 bookmarks. 

Potential reach of 5,045,699. The most revealing metric is the post type breakdown: 44.44 per cent of all posts — 88 out of 198 — are replies. Replies, in crisis terms, are not the shape of casual observation. They are the shape of argument, complaint, troubleshooting, defending, and escalating happening simultaneously across interconnected threads.

There is also a critical nuance in the sentiment that changes the communications strategy in a meaningful way. At post level, sentiment is 12 per cent negative. At impressions level, it shifts to 8 per cent negative, because high-follower accounts are more likely to post neutral or analytical commentary while the large volume of regular users generate the negative posts. 

This tells you that negative posts dominate emotionally and culturally even when impression-weighted sentiment appears more balanced, because regular-user complaints feel more authentic and relatable to other regular users than high-follower neutrality. Safaricom’s response cannot be aimed primarily at influencers and media. It must address the mass of ordinary customers who are living the problem and saying so publicly.

The Memes, The Influencers, & A Competitor’s Opportunistic Trolling 

No crisis analysis of this situation would be complete without acknowledging the dimension that data can measure in volume but not in cultural flavour: the memes. Within days of the launch, Kenya’s formidable community of content creators, influencers, and KOLs had turned My OneApp into reliable comedic material. 

The complaints arrived not just in frustrated prose but in creative formats — screenshots turned into punchlines, instructional videos mocking the airplane-mode-toggling workaround, and the kind of dry Kenyan wit that turns corporate embarrassment into national entertainment. 

These are not just jokes. In Kenya’s social media culture, humour is one of the most effective and most enduring forms of brand damage, because once a company becomes a punchline it has to work twice as hard to be taken seriously again.

Then, a competitor arrived. Six days after the launch began generating its most sustained criticism, Airtel Kenya’s official Twitter/X account posted: “With My Airtel App, you don’t need to switch anything off and on to just check balance. Just Vibes and Convenience 😎 #MyAirtelApp.” 

It does not mention Safaricom by name. It does not need to. Every Kenyan on the internet knew exactly what it was referencing, and the timing was immaculate. In a week where Safaricom had been struggling to find the right words, their main competitor landed a single tweet that said the thing everyone was thinking in twelve words. In a week where Safaricom had continued promoting a broken app for four days before acknowledging it was broken, being beaten at their own social media game was perhaps the most stinging indignity of all.

What Customers Are Actually Saying: Eight Complaint Clusters

Data dashboards tell you the scale of a crisis. Customer voices tell you its anatomy. When you aggregate the conversation across Twitter/X, Facebook, WhatsApp groups, technology journalism, regulatory statements, and the replies to the apology post itself, eight distinct complaint clusters emerge with enough consistency and specificity to constitute a clear and prioritized product brief.

1. The first and most emotionally charged cluster is the diaspora lockout. OneApp requires a live Safaricom cellular data connection for authentication — not Wi-Fi, not another network’s data, not a VPN. For Kenyans living and working abroad who depend on M-PESA to send money home for school fees, rent, medical bills, and family support, this is a digital severance from their financial lifeline. Diaspora users were told to use data roaming for verification — an extraordinary consumer rights failure for a company whose global reach depends entirely on the trust of its most mobile customers.
2. The second cluster is the SIM slot dependency. The app only works correctly when the Safaricom SIM is physically inserted in Slot 1 on a dual-SIM device. For users who keep their Safaricom SIM in Slot 2 — entirely normal behaviour for a dual-SIM consumer who uses another network as their primary data connection — the app simply does not work until they physically swap SIM cards. For the majority of Kenya’s digitally active consumers who are multi-SIM by default, this is not a minor technical constraint. It’s an approach that does not understand its own market.
3. The third cluster, and the one that has escalated into regulatory territory, is the forced migration without consent. Kenya’s leading consumer protection body has publicly described what happened as a “deploy first, fix later” approach with no credible evidence of user testing, feasibility study, or consumer impact assessment before rollout. The new app has been characterized as demonstrably worse than what it replaced — the old M-PESA App was lean, fast, and familiar, while OneApp is bloated, unstable, and intrusive. That verdict, from an established consumer watchdog, is not one Safaricom can afford to let stand as the defining public characterisation of their flagship product.
4. The fourth cluster is the biometrics regression. Biometric authentication settings are not consistently retained, requiring users to re-enable security features after every device restart. For a product whose marketing emphasises enhanced security, this is a particularly damaging contradiction that punishes the most security-conscious users most severely.
5. The fifth cluster is the cost-shifting complaint. Customers are asking a pointed and entirely reasonable question: why should they need to purchase Safaricom mobile data simply to access their own account on Safaricom’s own app? For users who have already migrated to competing networks for primary data — and there are millions of them, as the Communications Authority statistics confirm — this friction is not just an inconvenience. It is a tax on loyalty that actively accelerates the churn Safaricom is trying to prevent.
6. The sixth cluster is the lost favourites — years of saved paybills, till numbers, and frequent contacts wiped overnight with no migration, no recovery tool, and no advance warning. The phrase circulating in WhatsApp groups captures the emotional reality perfectly: they wiped my life.
7. The seventh cluster is older device incompatibility. Some older Android devices cannot complete the setup process at all. These are the least vocal users in the data and the least likely to be generating KOT memes. But they represent the most ethically serious dimension of the failure: customers with the least technological flexibility being silently excluded from their financial platform without acknowledgement.
8. The eighth cluster is the rollback demand — customers who are actively sideloading the deprecated M-PESA App from third-party mirrors, uninstalling OneApp entirely, and disabling automatic updates to prevent being pushed back. These are not passive complainers. These are customers making deliberate, technically effortful, active choices to reject the new product. The circulation of rollback guides across WhatsApp groups and technology forums is itself one of the most powerful and telling expressions of UGC in the entire crisis.

Reading The Reply Threads: What The Apology Post Actually Generated

The most diagnostic data in this entire crisis is not the metrics or the post-type breakdown. It is the content of the replies to Safaricom’s own apology post on Twitter/X, which I have analyzed as of this afternoon, and which paint a picture of customer sentiment that is simultaneously more specific, more human, and more commercially alarming than any aggregated data-driven insights. 

The apology is being actively rejected by a significant proportion of the people it was meant to reassure. The dominant sentiment in the thread is not gratitude for the acknowledgement, nor relief that help is coming. It is a combination of exhaustion, skepticism, and in some cases outright dismissal. 

Across the visible replies, multiple customers are explicitly saying they no longer want the apology — they want the old app back, or they want the new app to actually work, and they want that tonight, not in six months. Several are saying they are simply done. These are not the words of customers who are calibrating their response. These are customers who have reached a conclusion.

The diaspora and connectivity suffering in the thread is not abstract or hypothetical. One customer described travelling for roaming access every time their login expires at a remote work site and then losing the session when they return, burning significant fuel in the process, and asking for compensation. 

Another described waking up to find the M-PESA App simply gone and being unable to pay bills or fulfill financial commitments from abroad, having been given no choice about the migration and no warning it was coming. These are consumer protection cases dressed in the language of a social media reply thread.

The most humanly devastating account in the entire thread describes spending over three hours on the phone helping a 70-year-old family member navigate the new app login process — a family member who ultimately gave up and returned to a feature phone. There is more analytical content in that single account than in most technology reviews. 

A 70-year-old M-PESA user — the kind of customer who represents both M-PESA’s greatest social achievement and its most vulnerable demographic — spent three hours with family trying to navigate a new app and ultimately reverted to a device with none of its complexity. The Super App vision is aspirational and commercially sound. But it was built for a specific kind of user. The 70-year-old who was perfectly functional on the old M-PESA App is not that user. There are thousands if not millions of others like her.

The most actionable technical feedback in the thread can be summarized in three points that Safaricom’s product team should treat as a direct brief: replace the SIM card and Safaricom data requirements with OTP authentication; require OTP only when the app is used on a new device, not on every update; and — the three most condensed and most damning words in the thread — learn from banks. 

Kenya’s banking apps have managed OTP-based authentication over any network from anywhere in the world for years. This is a solved problem in every comparable fintech context globally. The idea that a Safaricom financial utility in 2026 requires a physical SIM in Slot 1 on a live Safaricom cellular connection is, as multiple customers expressed, backward thinking by a company that should know better.

The accountability demand in the threads is escalating beyond generic corporate responses. Customers want to know that the people responsible for the decision-making that produced this outcome are being held to the same standard the failure has imposed on ten million users. The sentiment is consistent: all the money, all the engineers, all the time — and this is what was shipped. Was there no testing? The question is not rhetorical.

The competitive sign in the thread is the most commercially alarming element of all. Airtel comparisons are being made spontaneously and organically, without any prompting from Airtel’s marketing. Multiple customers are independently arriving at the same observation: that the competitor’s app does the core things the old M-PESA App did, simply and reliably, without the friction, the SIM slot demands, or the Wi-Fi prohibitions. 

In one of the single replies in the entire thread, one customer responded to Safaricom’s formal apology not with acceptance or continued anger, but by publicly redirecting their gratitude to Airtel Kenya. That reply — Safaricom’s own apology post generating a competitor endorsement in its first thread — encapsulates the commercial stakes of this crisis more vividly than any data dashboard ever could.

Then there is the reply that validated, in nine words, everything I wrote about the psychology of Safaricom’s customer relationships in my Airtel Money piece last week. A customer described the dynamic as a monopoly treating its customers like a battered wife. Unprompted. In a reply to a corporate apology. That is not a social media post. That is a diagnosis.

Comparing This To Kenya’s Previous Major Social Media Crises

The Co-op Bank crisis of March 2022 remains the most directly instructive precedent. When a single customer with a handful of followers triggered a KOT firestorm over M-PESA deposit charges, the bank reversed the charges and announced refunds within hours — but the brand wound persisted. 

The lesson I drew then remains directly applicable now: money is one of the most sensitive topics for people from all walks of life, and if you plan to do anything that affects your customers financially, you must over-communicate before, during, and after. Silence reads as arrogance. Delayed acknowledgement reads as indifference. Neither is recoverable quickly.

The KFC chips crisis of 2022 offers the more optimistic lesson. KFC’s revelation that it sourced chips from Egypt rather than Kenya emerged accidentally from an innocuous supply shortage post. What made the KFC story ultimately instructive was the company’s decision to treat the crisis as product feedback rather than a PR problem. 

The subsequent work to identify Kenyan potato suppliers who could meet quality standards, the eventual local sourcing, and the public communication of that transformation through billboards and media two years later, represents exactly the approach that a product crisis should follow. 

The question for Safaricom is whether OneApp becomes a Co-op Bank story — forced capitulation that leaves a persistent brand wound — or a KFC story — genuine product reckoning that ultimately produces a stronger platform and a more trusted brand on the other side.

The Art Caffe and Chandarana crises add the dimension that explains the emotional intensity of the OneApp backlash. Both were driven by a perception that a brand did not respect its core Kenyan customer base — that it assumed loyalty without earning it. That same emotional scenario is visible across the OneApp conversation. The battered wife metaphor that appeared organically in the reply thread is not simply a complaint about an app. It is an articulation, in nine words, of the accumulated emotional cost of a relationship that has always been more dependent than it has been equal.

The Rollout Strategy Was A Fundamental Error

Safaricom chose what systems professionals call a Big Bang approach — the entire new system deployed to all users simultaneously as a forced auto-update, the old system replaced overnight for everyone with auto-update enabled. 

The parallel adoption approach — running the new app alongside the old one for a defined period, migrating users in deliberate cohorts, maintaining a clear fallback — would have served Safaricom’s customers and ultimately Safaricom’s brand infinitely better. 

This is not exotic technology practice. It is standard procedure for mission-critical systems. The same lesson had already been learned, at considerable public cost, from the Government’s SHA system rollout — a precedent that makes this choice of approach even harder to explain. 

The fact that Safaricom has confirmed it intends to decommission both the M-PESA App and the MySafaricom App within six months of the OneApp launch means the clock is now running on the parallel window. If OneApp is not genuinely reliable and feature-complete within that timeframe, the decommissioning will trigger a second crisis that will make the current one look manageable.

The Seven-Point Crisis Playbook Safaricom Needs To Execute Right Now

Every organisation faces risks on social media. The difference between a crisis that recovers and one that defines a brand permanently is almost always the quality and speed of the response across four phases: Readiness, Response, Reassurance, and Recovery.

The Readiness phase has passed and its failures are retrospective lessons. A systematic pre-launch risk assessment against Kenya’s real consumer behaviour — multi-SIM users, diaspora customers, elderly users, customers on non-Safaricom data — would have identified every failure mode that materialised. A pre-prepared holding statement would have been deployable within hours of the first complaints. A KOL and influencer mapping exercise would have enabled proactive engagement before the launch rather than reactive damage control after it.

The Response phase began in earnest today with the apology. The apology is competent in content — specific about who was affected, not deflective — but critically incomplete in two respects. As of this writing, it had not appeared on Instagram, LinkedIn, or any of Safaricom’s other social media platforms, which is inexplicable given that the diaspora and professional audiences concentrated on those platforms are among the most affected users. It contains no timeline — customers are told things are being fixed but not told when. In a product crisis, “we are fixing it” without a timeframe is not reassurance. It is a continuation of uncertainty.

The seven-point operational playbook for the Reassurance phase is this:

1. Eliminate the Wi-Fi authentication barrier as an emergency fix with a public date commitment — this single action resolves the diaspora crisis and the multi-SIM user crisis simultaneously, and the OTP-based authentication that would achieve it is a solved problem in every comparable fintech context. 
2. Restore or provide a fast import mechanism for saved favourites with explicit, honest communication about what can and cannot be migrated. 
3. Fix the statements reliability with proper error handling, cached fallbacks, and meaningful status messaging — timeouts in a financial utility are trust failures, not performance metrics. 
4. Extend the formal apology to Instagram, LinkedIn, and other social media within 24 hours. 
5. Publish a specific, dated “OneApp Stability Update” every week in plain language listing what has been fixed, what is in progress, and what comes next — this creates the public accountability loop the data tells you the conversation is demanding. 
6. Stop all promotional OneApp content until the basic product works reliably for the full user base — promotional posting during an unresolved crisis is not tone-deaf marketing, it is active brand damage. 
7. Stand up a real-time war room with live monitoring of both internal reliability metrics and external social media listening dashboards, because you cannot manage a crisis you cannot measure in real time.

Recovery — & What It Actually Requires

Recovery is not achieved through a marketing campaign celebrating improvements. It is achieved through the improvements themselves, communicated in plain language, followed by customers who voluntarily and publicly say the app now works. The fastest path to Recovery is a product that people use and recommend because it is genuinely good — not because they feel they have no realistic alternative.

Safaricom’s OneApp is currently trending in Kenya. That is visible in the Twitter/X sidebar of the very thread where these replies are being posted. A company that spent 25 years building Kenya’s most trusted brand is trending today not for a milestone, an achievement, or a celebration. It is trending because of an app. In the reply thread to its own apology, a customer drove 200 kilometres to access roaming, a grandmother returned to her feature phone, and someone responded to a corporate apology by publicly endorsing a competitor.

Safaricom has the resources, the engineering talent, and the strategic vision to make My OneApp genuinely world-class. What the past two weeks have demonstrated is that good infrastructure and good intentions are not sufficient on their own. They need to be matched by a product culture that places the customer — including the 70-year-old grandmother, the diaspora parent, the multi-SIM data-optimizing millennial, and the customer on an older Android device — at the centre of every design decision, not at the end of a press release.

A fully functional OneApp — fast, reliable, migrated, intuitive, globally accessible, and trusted — is the only statement that ultimately matters to the 40 million Kenyans whose daily financial lives run through M-PESA. They do not need Safaricom to be perfect. They need Safaricom to be reliably excellent at the one thing no competitor has (yet?) managed to take from them. Right now, that infrastructure is running on a product that is not yet good enough. The clock is ticking, the data is visible, the customers have spoken in terms that require no translation, and the competitive alternatives are a download away. What happens next is entirely up to Safaricom.