The strategy I’d approach
Revolut is excellent right up until something goes wrong.
Then it goes quiet.
The product is well loved. 4.7 on Trustpilot, 4.8 on Google Play. It also draws more ombudsman fraud complaints than any major UK bank. Both are true, and the gap between them is one journey nobody owns: what happens after money is frozen, blocked or stolen.
“A further £67,000 was stolen in the 23 minutes it took to reach the team that could freeze the account.”
BBC Panorama, October 2024 · a customer who lost £165,000
Four failures, one cause. A restriction arrives with no state and no end date. The only route to a person sits behind a bot. A fraud decision arrives without its reasoning. And growth is outrunning the navigation. The happy path has a product team and a roadmap. The recovery path has fraud, compliance and support, who are measured on risk and cost. So it never gets designed.
- 01 Make every restriction legible A status object in the app: what happened, what we can and cannot say, what happens next, and by when. Automatic, not on request. Day 60 · measured on repeat contacts per case
- 02 Build one fast lane for emergencies A fraud-in-progress route that skips the bot, triggered by the customer saying it, plus a freeze they control themselves. Day 90 · measured on time to human
- 03 Show the reasoning behind every decision A written rationale in plain language, naming the evidence used and what would have changed the answer. Quarter 2 · measured on the ombudsman uphold rate
Ranked by damage, not by ease. The fourth problem, navigation, is real and it is fourth: a cleaner home screen does not help someone whose salary is frozen.
The problem, in four parts
Revolut has two populations using one product. Ninety-three per cent of reviewers rate it four or five stars. The other seven per cent hit a wall and cannot get past it. Those are not the same journey with different luck. They are the happy path and the recovery path, and only one of them has been designed.
Ranked by damage to the customer
-
01
A restriction arrives with no state and no end date
Frozen accounts, blocked payments, transfers stuck pending, accounts closed with no reason given. The customer is not told what triggered it, what would clear it, or how long it takes. This is the biggest theme in every negative source I read.
What the user actually experiencesEVIDENCE “My payment has been blocked. I don’t know what the problem is.” · “My account blocked and Revolut is asking me to fund my balance” · a one-star review titled “Strong anti-fraud policy, but no protection for vulnerable customers”Your salary landed on Friday and the account locked on Saturday. Rent is due Tuesday. The app says your account is under review. It does not say by whom, for what, or until when.
-
02
The only route to a person sits behind a bot
Support is in the app only. No phone, no branch. Revolut’s own account says the assistant handles 75 per cent of questions effectively, with 2,000 agents behind it. For a card query that is good. For an active theft it is the whole story.
What the user actually experiencesEVIDENCE BBC Panorama, October 2024: £67,000 taken during a 23-minute wait · over 100 more customers contacted the BBC afterwards · a former employee said financial crime work “always played second fiddle to the desire to launch new products”You can see the money leaving. You type “I am being scammed” and the bot offers you an article. By the time a person reads it, the account is empty and the conversation turns into a claim.
-
03
A fraud decision arrives without its reasoning
Claims are refused, the reason is thin, and the ombudsman then finds for the customer often enough that distrust is rational. Revolut is not the largest of the three firms below. It draws the most complaints.
What the user actually experiencesEVIDENCE FOS 2024 APP fraud: 3,242 Revolut, 2,344 Monzo, 1,704 Barclays · other fraud: 2,631 Revolut, 1,810 Monzo, 1,461 Barclays · Jan to Aug 2025 uphold rates 30% and 37%Weeks after losing your savings, a message says the payments were authorised by you, so there is no refund. Nothing says what evidence was used. So you go to the ombudsman, and about a third of the time they agree with you.
-
04
Growth is outrunning the navigation
In one year Revolut added ATMs, mortgages, mobile plans, lending, a teen product and RevPoints across 36 countries. A published heuristic review put a plain wire transfer at six screens against three in a traditional banking app, with no clear head navigation.
EVIDENCE third-party heuristic review, May 2025 · Revolut’s own 2025 annual report · I have not audited the app myself, and the method section says so again
Those four are the problem. Three of them share one cause: the recovery path was never designed, because designing it is nobody’s objective. That is what the strategy attacks, in that order.
The strategy
Three moves, in order, each answering one of the problems above. I would treat the recovery path as one product surface rather than a support backlog: everything a customer experiences between “something is wrong” and “it is resolved”.
Move 01 · answers problem 01
Make every restriction legible
A status object shown in the app for every restriction: what happened, what we can say, what we cannot say and why, what happens next, and by when. Updated automatically rather than when the customer asks. Plus a named severity ladder, so a stolen salary and a duplicate card charge stop sharing a queue.
SHIPS → live on the two highest-volume restriction types by day 60. Measured on repeat contacts per restriction case.Move 02 · answers problem 02
Build one fast lane for emergencies
A fraud-in-progress route that skips the bot entirely, triggered by the customer saying it rather than by the bot deciding it. Plus a one-tap freeze the customer controls themselves, so nobody waits 23 minutes to stop their own account. The bot stays exactly as it is for everything else.
SHIPS → live by day 90. Measured on time to human, and on money moved after the first customer report.Move 03 · answers problem 03
Show the reasoning behind every decision
A written rationale on every fraud outcome, in the app, in plain language, naming the evidence used and what would have changed the answer. The ombudsman file is read monthly as research rather than as legal risk, and coded by what the first decision missed.
SHIPS → piloted in quarter 2. Measured on the ombudsman uphold rate, as the honest lagging indicator.Why this order, and the number I would put on it
Silence is cheaper to fix than speed, so it goes first. Nobody currently owns the question “how long has this customer been restricted without being told anything”. I would build that measure in the first month, split by trigger, and report it beside growth. If a review genuinely needs five days, fine. Five days of silence is a different thing, and it is the one that costs you the account.
The three arguments I would have to win
None of these moves is obvious, because each runs into a real constraint held by a team who is right to hold it. Pretending otherwise is how product owners get politely ignored.
Tell them more vs. tip off criminals
Financial crime rules limit what a bank may disclose, and a detailed explanation is a manual for evading the next check.
How I resolve it: separate what we cannot say from the fact that a limit exists. Nothing prevents telling a customer that a review is running, that we are legally limited on the detail, and when we will next update them. Almost all of the anger is about the silence, not the secrecy.
Automation vs. a person now
The bot resolves 75 per cent of contacts and is why support scales to 80 million customers. The 25 per cent it cannot solve includes every case where minutes cost thousands.
How I resolve it: stop treating them as one queue. A declared emergency skips the bot and lands on a human with freeze authority. Triggered by the customer, because a model that gates the emergency route is the failure it is meant to prevent.
Ship the next product vs. fix the last one
Growth is the strategy, and a million new customers every 17 days is not an accident. A former employee told Panorama that financial crime work always came second to launches.
How I resolve it: make the recovery path a product with a roadmap rather than a cost line. It then competes for engineering on the same terms as everything else, and stops being the thing that only gets attention after a documentary.
Why this is available to Revolut specifically
- 01
Nobody in this category has claimed the worst day
Every neobank competes on the happy path, and the ombudsman numbers say none of them has solved the other one. The first to make a restriction legible end to end takes a position that does not depend on being cheapest or fastest.
- 02
The banking licence changes what is at stake
As of March 2026 these are protected deposits and this is a main account. A frozen holiday card is an inconvenience. A frozen salary is a different product failure, and the customer base is shifting from one to the other right now.
- 03
The staffing is already there
More than a third of a 10,000-person workforce sits in fraud prevention, with over 2,000 support agents behind the bot. This is not a hiring problem. It is a routing and communication problem, which is cheaper and faster to fix.
The evidence
Where the four problems came from. Two things up front: reviews lean angry by nature, and I have no access to your own research, which almost certainly says more than this does.
Four scoreboards, and they disagree on purpose
The app store scores are excellent and I am not going to pretend otherwise. The complaint data is the worst of any major UK bank. Both are true, and the gap between them is the whole finding.
Seven per cent of 454,000 is roughly 32,000 one-star reviews. Most companies would take a 4.7 without blinking. Almost none of them also have thirty-two thousand people writing the worst review the platform allows.
The complaints, sorted
Fourteen, collapsed by default. Open a bucket, or filter to one. Where something comes from a source I could not verify properly, I have said so on the item rather than in a footnote.
Critical 5Costs the customer money or the relationship
The most common theme in every negative source. Salaries locked during a compliance review with no stated cause and no stated end. The money is not gone, which is what makes it hard to argue about, and it is also completely unavailable on the day rent is due.
Support lives in the app behind a chatbot. During an active theft that is the difference between a freeze and a claim. Panorama documented £67,000 taken inside one 23-minute wait.
The decision arrives, the reason does not. The ombudsman then upheld 30 per cent of APP fraud complaints and 37 per cent of other fraud complaints against Revolut between January and August 2025.
Long-standing accounts terminated without explanation. There are real legal limits on what a bank may disclose. There is no legal limit on saying clearly that a limit exists, which is what is missing.
Verbatim from a review: “My payment has been blocked. I don’t know what the problem is.” The block may well be correct. The sentence after it is the product failure.
Serious 5Erodes confidence and generates avoidable contact
Revolut states the assistant handles 75 per cent of questions effectively. The complaint is about the other 25 per cent, where the same scripted lines return and there is no visible escalation. A bot that is right three times out of four still needs an exit.
Pending is a state with no duration attached. Customers cannot tell a two-hour check from a two-week one, so they treat every hold as the worst case and contact support to find out.
Document and selfie checks that fail without saying which part failed. The retry loop is the same screen with the same instruction, which is the definition of a journey with no state.
The Panorama case turned on a caller claiming to be from the bank. Revolut has since shipped an in-app banner that confirms whether a call is genuine, which is a good fix. It only works if the customer thinks to check.
The Panorama customer was on a business account. A frozen business account stops payroll, not just a weekend. The severity is different and the route to a person is the same.
Neutral 4Real, but lower impact or outside this remit
A published heuristic review counts more calls to action than it can list and puts a wire transfer at six screens against three elsewhere. Real, and fourth, because it costs seconds rather than salaries.
Raised often, largely a disclosure and timing problem rather than a pricing one, and mostly already explained in the app. Included so the reading is honest rather than selective.
Ordinary admin friction that every card issuer draws. It sets the baseline expectation a customer brings to a real problem, which is why it is here at all.
One site puts the one to three star themes at roughly 30 per cent frozen accounts, 24 per cent chatbot-only support, 18 per cent sudden closure, 16 per cent refused refunds and 12 per cent holds and fees. It publishes no sample size and no dates, so I have treated it as indicative and built nothing on it.
How that compares
The honest finding: Revolut beats the traditional banks on almost everything except the moment it matters most. The high street is slower and better reviewed by nobody, and it still has a phone number a frightened person can dial.
Where the customer actually feels the difference
The read: Revolut just became a fully licensed UK bank holding deposits protected to £120,000. The people arriving now are moving their main account, not their holiday money. The cost of a bad recovery path goes up the moment it is someone’s salary rather than their spending card.
The recovery path
What I would actually draw. Everything a customer experiences between “something is wrong” and “it is resolved”, blueprinted across the customer, the app and the teams behind it, with the six places it currently breaks marked on it.
The path as the customer experiences it
Six stages. The blue band is what the customer can see. The line underneath is what is really happening. The gap between them is the whole problem, and every magenta mark is a moment where somebody opens the chat.
The six break points, and what I would ship at each
- 01
Before it happens: nobody knows this is possible
Customers do not know a compliance review exists, what triggers one, or that it can hold a balance. So the first time they meet it, it reads as theft rather than process.
SHIP → a short, honest explainer in onboarding and in the help centre: this can happen, here is roughly why, here is what we will do. - 02
The trigger: a banner with no content
“Your account is under review” is a state without a cause, an owner or an end. It is technically accurate and practically indistinguishable from nothing.
SHIP → the restriction status object: what happened, what we can say, what we cannot and why, what happens next, and by when. - 03
The review: the long silence
The longest stage and the emptiest. Internally it is busy the whole time. Externally it looks identical to being ignored, so the customer contacts support repeatedly, which slows the queue that is holding their case.
SHIP → automatic updates on change rather than on request, with an honest expected time. This is the single highest-value change on the path. - 04
Contact: one queue for everything
A stolen salary and a duplicate card charge enter the same funnel. Panorama documented £67,000 taken inside a 23-minute wait. That is not a staffing failure, it is a routing failure.
SHIP → a named severity ladder and a fraud-in-progress route that skips the bot, plus a one-tap freeze the customer controls themselves. - 05
The decision: an outcome with no reasoning
The ombudsman upheld 30 per cent of APP fraud complaints and 37 per cent of other fraud complaints against Revolut in the first eight months of 2025. A decision that cannot be understood cannot be accepted, so it becomes a case.
SHIP → a written rationale in the app: what we found, what we relied on, and what would have changed the answer. - 06
After: the relationship is never repaired
Once a case closes, nothing happens. The customer who was frozen for four days and then cleared gets no acknowledgement, and quietly moves their salary elsewhere. That churn never appears next to the restriction that caused it.
SHIP → a close-out message that says what happened, plus a retention measure that joins churn back to the restriction that preceded it.
Accessibility
This is where I have the most direct experience, and also where I have to be careful about what I claim from the outside.
What I could and could not check
The app sits behind authentication and the recovery path only exists once something has gone wrong, so I have run no scan and no audit. Nothing here is a finding about Revolut’s accessibility. What follows is what the regulation requires and what I would do about it.
Where the obligation lands
The European Accessibility Act has applied since June 2025 and banking services are explicitly in scope. The expectation is WCAG 2.2 Level AA across the service, including the parts customers only meet on a bad day, which are usually the least tested.
Why it belongs in this role
Someone using a screen reader while their account is frozen meets every failure above, and then some. I hold the Web Accessibility Specialist certification and have audited against WCAG 2.2 since 2022. The useful skill is turning a standard into a short list of things to start with on Monday.
Four independent sources, with different populations and different motives, name the same thing: not the decision, the not knowing. That agreement is the strongest signal in this document, and it means the first month does not need a discovery phase to know where to start.
The plan
Everything above is desk research, and desk research is a hypothesis generator rather than an answer. This is how I would actually run it: the first ninety days, what ships when, and how you would know whether any of it worked.
The framing matters more than the method. The obvious brief is “reduce complaints”, and that brief is a trap, because the fastest way to reduce complaints is to make complaining harder. The real brief is to separate the time a review genuinely needs from the silence around it, and to attack the second one hard.
Find the real shape of the recovery path
Internal and secondary first. It is cheaper, faster, and it tells me what to ask customers.
- Pull every restriction, block and fraud claim from the last twelve months and map them by trigger, duration and outcome, so I know which triggers produce the longest silences
- Sit with fraud, financial crime and legal and write down what we are actually permitted to tell a customer. Most of this gap is legal constraint plus habit, and the two get confused
- Read 500 one-star reviews and 200 ombudsman decisions, coded by cause. Cheap evidence before expensive evidence, always
- Read the existing research. A company this size has almost certainly answered part of this already, and re-running it would waste a month
You get: a ranked map of the recovery path with volume, duration and cost against each failure point, and the five that account for most of the damage. Roughly 20 pages, and every team can use it immediately.
Ship the restriction status object
The cheapest win here is not faster resolution. It is telling people what is happening while it happens.
- A status object for every restriction: what happened, what we can say, what we cannot say and why, what happens next, and by when. Shown in the app and updated automatically
- An expected time, and a commitment to update it. An honest 48 hours beats a silent six
- A named severity ladder, so a stolen salary and a duplicate card charge stop sharing a queue
- Interviews with twelve customers who were frozen and twelve who were defrauded, half reimbursed and half not, to check the wording lands before it ships to millions
You get: the status object live on the two highest-volume restriction types, with a measurable target on repeat contacts per case.
Ship the emergency route
A concept that does not ship is a nice deck. Landing it is part of the method, not the afterthought.
- A fraud-in-progress route that skips the bot entirely, triggered by the customer saying it rather than by the bot deciding it
- A one-tap freeze the customer controls themselves, so nobody waits 23 minutes to stop their own account
- A written rationale on every fraud decision, in plain language, with the evidence used and what would have changed the answer
- The internal view of the same state, so an agent can answer “where is my case” without chasing it themselves
You get: the emergency route live, measured on time to human and on money moved after the first customer report. Two shipped changes inside ninety days.
The method mix, and why each one is there
Four methods. Each covers a specific weakness in the others, and naming the weakness is the part that usually gets skipped.
The whole path across customer, app and the teams behind it, with every handoff named. The only method that shows you where the silence comes from.
Shows the process as designed, not always as performed. Needs walking, not just workshopping.Every restriction joined to its contacts, its duration and its outcome. Tells you how often, where, and at what cost, at full scale.
Blind to the customer who never messaged and quietly moved their salary elsewhere.Customers who were frozen and cleared, frozen and closed, defrauded and reimbursed, defrauded and refused. Tells you what they believed was happening and what they did next.
Small n and reconstructive memory. Explains the numbers, cannot replace them.The one session that decides how far this can go. Writing down the real disclosure boundary, rather than the cautious version everyone assumes.
Produces consensus, which is not the same as truth. Needs a lawyer in the room, not a summary of one.The hypotheses I would test first
Four, each with the signal that would confirm it, the method that would test it, and what would have to move for a fix to count. This is the table I would want to be wrong about by day thirty.
| Hypothesised driver | What you would see | How I would test it | What has to move |
|---|---|---|---|
| Silence, not slowness | Contact volume spiking in the stages with no customer-facing update, not in the longest stage. | Restriction records joined to contact records, plus depth interviews. | Repeat contacts per restriction case. |
| The bot is the bottleneck only in emergencies | Satisfaction fine on enquiries, catastrophic on fraud-in-progress. | Split the contact data by declared intent and outcome severity. | Time to human on a declared emergency. |
| First fraud decisions are under-evidenced | Ombudsman overturns clustering on a small number of case types. | Code 200 published decisions by what the first decision missed. | Uphold rate, as the lagging indicator. |
| Restrictions drive silent churn | Customers cleared after a freeze reducing their balance and moving their salary. | Retention analysis joined to restriction history. | Balance and salary retained ninety days after a restriction closes. |
Four ongoing practices
A project answers one question. These are the loops that make the second question cheaper than the first, and that let the work continue without everything routing through me.
Practice 01
The blueprint stays alive
A blueprint that is drawn once and framed is decoration. This one gets revisited every quarter with the people who run the process, so it stays a description of reality rather than of a workshop held in March.
YOU GET → one shared picture of the path that product, fraud, legal and engineering all argue from.Practice 02
The ombudsman file, read as research
Every overturned decision read and coded by what the first decision missed, then fed back into the fraud rules and the wording. It is currently treated as legal risk. It is the cheapest labelled dataset we will ever get.
YOU GET → a monthly list of specific, evidenced first-decision failures instead of an annual embarrassment.Practice 03
The silence measure
How long a customer has been restricted without being told anything, split by restriction type. Nobody currently owns this number. Once it exists it changes what people argue about, because it separates the time a review needs from the silence around it.
YOU GET → a number that makes the case for this work without me having to make it.Practice 04
Coaching, not managing
Your post asks for someone who has managed people. I do it by critique against a written standard, so a junior can win the argument on the evidence, and by pairing on the hard half of a problem while leaving the work in their hands.
YOU GET → product decisions made in rooms I am not in, which is the only version that scales.How you would know it is working
Four measures, reported weekly, plus the ombudsman uphold rate as the slow one. The first is the one I would insist on, because it is the only one that tests the actual argument.
- 01
Time restricted without being told anything
Split by trigger, reported beside growth. It separates the time a review genuinely needs from the silence around it. If this does not fall, nothing else on this page mattered.
- 02
Time to human on a declared emergency
From the moment a customer says they are being defrauded to the moment a person with freeze authority reads it. Twenty-three minutes is the number in the public record. I would want a target and a chart, not a promise.
- 03
Money moved after the first customer report
The measure that turns support latency into pounds. It is also the one that will get engineering time approved, because it converts an experience argument into a loss argument.
- 04
Balance and salary retained ninety days after a restriction closes
The measure nobody keeps. A customer who was frozen, cleared and never trusted us again does not complain. They just stop being the main account, and no funnel metric will show you that.
One measure I would deliberately not use: satisfaction on the recovery path. People whose money was stolen do not give good scores, and chasing that number pushes a team to avoid the hard cases. I would rather be judged on time, money and retention, which are harder to argue with.
The thing I actually care about
I have spent six years turning research into things teams can act on the same week, in two-hour sessions with product owners, developers and analysts who did not commission the work and did not have to agree with it. What I care about is not the blueprint. It is the moment someone changes their mind because they finally saw what a person was actually trying to do. On this path that person is watching their savings leave and typing into a chat window, which makes it matter more than usual.
Method, and what to distrust
Sources used
- Review corpora: Trustpilot for revolut.com, Google Play for the main app, and PissedConsumer.
- Which? analysis of Financial Ombudsman Service fraud complaint data for 2024 and for January to August 2025.
- BBC Panorama, October 2024, and the follow-up reporting on customers who contacted the BBC afterwards.
- Revolut’s own published material: the 2025 annual report, the customer service explainer, the UK banking licence post, and the customer numbers post.
- Role context: your Product Owner (UX) advert, and what it says about ownership, data and design standards.
- Frameworks: Doblin’s Ten Types of Innovation, and Roger Martin on integrative thinking.
Known biases and limits
- The good scores are real and I have not discounted them. A 4.7 across 454,000 reviews and a 4.8 across 4.07 million ratings are not noise. Nothing on this page argues the core product is weak.
- Trustpilot self-selects at both ends. It measures how motivated people are to record an experience. I have used the one-star volume as a count of unresolved cases, not as a satisfaction score.
- I could not read a clean run of recent one-star reviews. Trustpilot’s one-star filter would not render for me and the App Store review page rate-limited me. So I have the distribution and a small number of verbatim quotes, not a coded sample.
- One third-party review analysis is unverifiable. The site giving theme percentages publishes no sample size and no dates. It is marked as indicative on the page and nothing rests on it.
- The regulator no longer publishes per-firm reimbursement rates. The last firm-level figure I can cite is from 2023: £756 of APP scam money per £1 million received. I have not used it as a current number.
- I have not audited the app. No scan, no walkthrough, no assistive-technology testing. The interface points are other people’s published analysis and are labelled as such.
- This is outside-in work. No internal data, no roadmap, no constraints, and no idea what your teams have already found. Two weeks inside would sharpen or kill most of it, and the plan is written on that assumption.
Sources
- Trustpilot: Revolut
- Google Play: Revolut
- PissedConsumer: Revolut reviews
- Which?: Revolut still the worst UK firm for fraud complaints
- Which?: response to the BBC Panorama investigation
- Good Money Guide: Revolut fraud response under the Panorama spotlight
- The Paypers: over 100 customers contact the BBC over Revolut scams
- Revolut: how our customer service works
- Revolut: is Revolut a fully licensed bank in the UK
- Revolut: annual report 2025
- Nimble: dissecting the app, Revolut
- Doblin: Ten Types of Innovation
- Roger Martin: The Opposable Mind (integrative thinking)
What you’ll be doing
“Turn complex problems into simple, scalable solutions.”
That sentence is why I applied.
I have a drawing about that. It is on every site I build and it has been the shape of my working life for twelve years: everything at once on the left, one line a team can build on the right. Your post describes the same move in different words.
And I am happiest at the left-hand end. Most people are drawn to the clean line. I like the knot: the contradictions, the four teams who each own one piece, the evidence that refuses to agree with itself. Below, each of your six duties with what I bring and what you would actually have in your hands.
What you’ll be doing
Your six bullets, each with what I bring and a deliverable with a date on it. I have kept your wording so you can check I have answered the actual question.
What you have by month twelve
Before the detail, the whole thing on one line each. This is what I would expect to be held to.
| When | What lands | How you check it |
|---|---|---|
| Days 1–30 | A ranked map of the recovery path: every restriction, block and fraud claim from twelve months, sized by volume, duration and cost. | One picture that fraud, legal, support and engineering all recognise as true. |
| Day 30 | The silence measure built and reported: how long customers are restricted without being told anything, by trigger. | A number that did not exist before, sitting beside growth in the weekly review. |
| Day 60 | Restriction status object live on the two highest-volume restriction types, plus a named severity ladder. | Repeat contacts per restriction case falling. Customers can say what is happening to them. |
| Day 90 | Fraud-in-progress route that skips the bot, and a customer-controlled one-tap freeze. | Time to human on a declared emergency, and money moved after the first report. |
| Quarter 2 | Written rationale on every fraud decision, in plain language, with the evidence used. | Ombudsman uphold rate, tracked as the lagging indicator. |
| Throughout | Weekly critique with written outcomes, pairing, and a one-page product direction per problem area. | Product calls being made in rooms I am not in. |
How my time splits
Commitments only mean something if they total one hundred. This is my whole capacity as a product owner who stays hands-on, and what happens when a launch is in flight.
Steady state
45 · 25 · 20 · 10 = 100
A fifth on alignment is deliberate. In a company this size, and on a path that crosses four functions, alignment is not overhead around the work. It is most of how the work actually happens.
Launch in flight
55 · 20 · 20 · 5 = 100
Delivery takes the load and I say so in advance rather than quietly missing something else. Coaching is the first thing squeezed and the first thing I put back.
-
01
Setting your team’s goals, success metrics and roadmap
The duty. Set goals, success metrics and a roadmap that align with company goals.
What I bring. The four measures on the plan tab are the actual answer: time restricted without being told anything, time to human on a declared emergency, money moved after the first report, and balance retained ninety days after a restriction closes. None of them exists today. Each one is defensible in a room with finance in it, because three of the four convert to pounds.
YOU GET → a measure that did not exist before, reported weekly from day 30, and a roadmap that argues from it. -
02
Driving maximum impact from data, market research and strategy
The duty. Drive maximum impact based on data analysis, market research and company strategy.
What I bring. This whole page is the sample. Four independent scoreboards, ombudsman data across two years, one broadcast investigation, your own annual report, and a competitor read, done from the outside with no access. My Baymard reports run to fifty or more recommendations and every one is ranked by transaction impact rather than by how annoying it is.
YOU GET → a ranked map of the recovery path with volume, duration and cost against each failure point, in the first month. -
03
Completely owning the product, ideation to maintenance
The duty. Own your team’s products from ideation through development to launch, growth and maintenance.
What I bring. At Allan Gray I designed the private investor platform from nothing and built the practice around it. At Luno I championed a full product redesign and acted as product manager across several agile teams. Every Baymard engagement is mine end to end: scope, method, execution, analysis, delivery, then the implementation call with the team that has to build it.
YOU GET → two shipped changes inside ninety days, not a strategy deck about them. -
04
Working with Design and UX to define the customer journey
The duty. Work closely with Design and UX teams to define the customer journey and create an amazing user experience.
What I bring. I am a designer who moved into product, which means I can write the brief in the language the design team already uses and I will not hand them a spec with the interesting decisions already made. The blueprint on the recovery path tab is the artefact I would bring to that first session, and it is deliberately unfinished in the two places where I do not yet know the answer.
YOU GET → a service blueprint of the recovery path that design, fraud and engineering all argue from, kept alive quarterly. -
05
Liaising with Engineering for effective delivery
The duty. Liaise with Engineering to ensure effective delivery of your product.
What I bring. A front-end background that never went away, HTML, CSS and PHP, and enough current curiosity to tell when a technical constraint is real and when it is a preference. That is most of what makes a product owner bearable to work with, because it stops the spec being something nobody can build. It also means I write acceptance criteria that survive a sprint.
YOU GET → specs an engineer does not have to translate, and fewer of the meetings that exist only because the brief was vague. -
06
Partnering with Marketing, Legal and Strategy
The duty. Partner with other stakeholders, such as Marketing, Legal and Strategy teams.
What I bring. On this particular product, Legal is not a stakeholder to inform, it is the constraint that decides how far the work can go. The single most valuable session in my first month is the one where legal and financial crime write down the real disclosure boundary rather than the cautious version everyone assumes. Six years of delivering findings to people who did not commission the work and were free to disagree is the training for that.
YOU GET → 20% of capacity protected for alignment, budgeted rather than borrowed, and a written disclosure boundary everyone can build against.
Three things you would otherwise have to ask me
Location, and the honest problem. Your post lists offices in Dubai, Dublin, London and Madrid, and remote in Germany, Ireland, Portugal, Porto, Romania, Spain and the UAE, plus one more that was cut off. I live in Dieren in Gelderland, and the Netherlands is not visibly on that list. I would rather raise that in the first line than have it found in week three. If the Netherlands is the missing entry, this is straightforward. If it is not, I am open to talking about a European hub and regular travel, and I would want to know before either of us spends more time.
Title. I have not held a role called Product Owner. I have done the work: product manager across several agile teams at Luno, and full ownership of every Baymard engagement end to end. If the title matters more than the track record, that is a fair thing to decide early.
Working language. English throughout, which your post confirms. I also have some Dutch and am actively learning, helped by growing up South African with Afrikaans around me, which comes from Dutch.
What you’ll need
You asked for three years in finance.
Here is the actual count.
Ten requirements on your list plus one nice to have. Eight are a straight yes, two are honest works in progress, and all of them are answered line by line below with the evidence attached rather than the adjective.
What you’ll need
Your list, line by line. Most are a yes. Two are honest works in progress, and both are in the table rather than hidden after it.
The list, line by line
Read the two that are not a straight yes first, because they are the ones that should decide whether this is worth your time.
| Requirement | What I hold | The evidence |
|---|---|---|
| 3+ years at a high-growth tech company or in banking and finance areas, including payments, investments and cryptocurrency | Yes | Two and a half years at Luno, a crypto exchange, leading design across onboarding, payments and security while the team grew from two designers to seven. Nearly two years at Allan Gray on the private investor platform. Six years at Baymard consulting to ecommerce and SaaS clients across Europe and the US |
| 2+ years of proven success leading consumer-facing mobile products | Yes | Luno is a consumer mobile product and I led design strategy across it. Research, journey maps, prototypes and visual design, plus a full product redesign I championed and then helped ship |
| 1+ years of experience managing people | Yes, with a caveat | I sourced and built the design team at Allan Gray as the first designer, including hiring the Head of Product Design, and helped grow Luno to seven. I have never held a title with manager in it: see the note below the table |
| Outstanding communication skills | Yes | Two-hour deep dives with product owners, developers, analysts and Director-level stakeholders who did not commission the work and were free to disagree. This page is the writing sample |
| An eye for great UX and UI, and impeccable understanding of design best practice | Yes | Sixteen years in digital and six of them auditing other people’s products for a living. Fifty or more recommendations per Baymard report, each with a mockup. Web Accessibility Specialist, auditing against WCAG 2.2 since 2022 |
| High analytical acumen with business sense guiding ROI-based prioritisation | Yes | Every Baymard report is ranked by transaction impact, not by severity of annoyance, and delivered with the fix placed in a component so it ships once. The plan on this site is prioritised the same way |
| Formulating and driving a focused product vision, owning the narrative with leadership, customers and partners | Yes | I championed the Luno redesign to leadership and then carried it. At Baymard I present the argument to teams who have to be persuaded rather than told. I thrive in presentation mode and I enjoy the writing that goes with it |
| Experience leading product development and launch from scratch | Yes | The Allan Gray private investor platform from nothing, and digital government services at the Western Cape Government covering document services and home and business applications |
| Insight and measurable results under ambiguity, in a fast-paced environment | Yes | This is the part I like most. The squiggle on this site is the argument, and the whole page was produced from the outside with no internal access |
| An entrepreneurial drive to never settle | Yes | I moved continents in 2022, six months pregnant, and rebuilt a career inside institutions I did not yet understand. I also build a site like this one for every role I care about |
| P&L ownership and commercial targets | Partly | You did not ask for this and I am adding it, because it is the honest gap. I have prioritised by impact and argued for spend, but I have never owned a revenue line: see the note below |
| Nice to have: a bachelor’s degree from a top university | Yes | BA Interaction Design, Vega, 2006 to 2009. MPhil in Inclusive Innovation for Information Technology, UCT Graduate School of Business, 2014 to 2015, ranked first in Africa |
I have never held a manager title, on purpose
For most of my career I steered towards depth of craft rather than headcount: being the person who does the work, sets the bar and brings others up to it, rather than the person who owns the org chart. I have still built teams from the first hire, hired designers, and hired two managers who went on to lead product design functions.
Your post asks for one year of managing people. I have done more than that in substance and none of it in title. If the requirement is line management with performance reviews attached, say so and I will tell you honestly that it would be my first time.
I have not owned a revenue line
ROI-based prioritisation I can do and have done, because it is how every Baymard report is ranked. Owning a number that finance holds me to is different, and I have not done that.
What I would do about it is exactly what is on the plan tab: pick measures that convert to pounds. Money moved after the first customer report, and balance retained ninety days after a restriction closes. Both are commercial arguments rather than experience ones, and both would put me in the room where that muscle gets built.
Why this role and not another product role
I have spent six years telling companies what is wrong with their products and then watching someone else decide whether to fix it. This role is the other side of that. It owns the thing end to end, it sits in payments and fraud where I have actually worked, and the problem underneath it is one I would be genuinely glad to spend a year on: making sure that the worst day a customer has with Revolut is a day they can understand.
Get to know the human behind all this
Irani
Sangham
Human on planet Earth, mother, and a terrible pun pusher.
Get to know the human behind all this
A little about me that has nothing to do with design.
South Africa to the Netherlands, via a great deal of learning
I was born and raised in South Africa. On weekends I was either hiking the wild mountains or surfing in the ocean.
In 2022 my husband and I moved to the Netherlands. He got a job here, so we decided to give it a try, and I was six months pregnant with my son at the time. Navigating a country move while pregnant is no small feat.
I have two children, and life since has been full of learning new things and upskilling incredibly quickly, in a completely new country, on the literal other side of the planet from my familiar environment and support system. It has been the most difficult and challenging time, but I have enjoyed making every day just a little bit easier by consistently trying my best, and those skills are directly transferable to this role.
I know what it is to be new to a system
I moved to the Netherlands in 2022, six months pregnant, and set up a life inside institutions I did not yet understand. Banks, registrations, forms, and a lot of waiting without knowing what was happening. Afrikaans gave me a head start on the language, which turns out to help far less than you would hope when the process itself is the confusing part. Opening accounts in a country whose rules I could not read is the closest I have come to being the person on the other side of a system designed by people who already know how it works. It is the whole reason the recovery path is the thing I want to work on.
Fun to have around, allegedly
I have been told I am very fun to have around and that I have a knack for making people laugh. I do love a good and bad joke, and I have disastrous puns, but I am not sorry about that.
I chose depth of craft over headcount, on purpose
You will notice I have not held roles titled manager or director. That is deliberate, and it is not a gap. For most of my career I steered towards depth of craft rather than headcount: being the person who does the work, sets the bar and brings others up to it, rather than the person who owns the org chart. Six years ago I made it explicit and moved into a consultative role, where the whole job is holding a standard and persuading people who do not report to me to meet it.
I have still built teams from the first hire, hired designers, and hired two managers who went on to lead product design. Your post asks for a year of managing people and I have done more than that in substance, just never with the title. Influence without authority is the only way I have ever worked, and on a path that crosses fraud, legal, support and engineering, it is the version that actually gets things shipped.
Some things I like
Competitive swimming
As a child my goal was the Olympics. I trained four hours a day and represented my country on numerous occasions, from age eight to fourteen.
Shaolin Kung Fu
From fourteen, I trained for fifteen years. I taught it, and competed for South Africa.
Ambidextrous, and juggling
Balls, and life. Both ongoing.
Coffee, and sourdough
A long-standing coffee nerd, and recently obsessed with sourdough and its many applications.
The growing brain
A lot of my time goes into toddler activities and investigations at the moment. The growing brain is fascinating to me.
Mountains and ocean
Hiking and surfing were the shape of every South African weekend, and the habit has not left.
Testimonials
She earned the complete trust of my entire team.
Lauren Hunt · on working together at Luno
Four colleagues and two clients, on what it is actually like when I am in the room.
Irani has had a massive impact on my career as a product designer. She has always made time to give me advice, resources and guidance to further my career, and continues to do so. She inspires through being extremely knowledgeable and capable while remaining immensely humble. She earned the complete trust of my entire team at Luno, and I would jump at the opportunity to work with her again.
Irani is very good at holding complexity and clarity, opposing points of tension, in her head simultaneously. She can empathise with the user and understand the problem while recommending a simple solution for the team to implement.
Irani, thank you for making an incomparable difference. You are an integral part of the team and possess a quiet composure, low-key witticism (and a vast supply of puns), and a warmth that makes it a pleasure to work with you. Not only that, but over time you have improved the work flow of the team by continually coming up with clever, innovative ideas.
Kudos going straight to Irani for organising an awesome summit. Practical, insightful, kept to schedule and beautifully facilitated.
Irani has made accessibility tangible and actionable for us. We have a clear list of where to start, removing the confusion and overwhelm. Thank you, Irani.
Irani has an exceptional ability to hold a room and get people to be productive. I felt like I had ideas flowing, and that any idea I put down was a good contribution. Great facilitation and run-through of our product improvement possibilities.