Telecom · Executive summary
What should management do?
The three sections below are deliberately kept apart. What is observed comes from a primary source and can be checked. What is inferred is our reading of that evidence and could be wrong. What is recommended is a judgement about what to do, and carries a test that would show whether it was right.
Observed
primary sources, checkable OBS The base is growing again, but has not recovered its peak 839,615 mobile subscriptions at May 2026, against 907,763 in September 2017. Nine years, and the market is still 7% below where it was. Source: CAM monthly statistics, 173 months.
OBS Mobile revenue is flat while the base grows Combined mobile service revenue was MVR 3160m in 2025 against MVR 3156m in 2024, on an average base 5.2% larger. Source: Both operators' quarterly income statements.
OBS Revenue per subscription fell 4.9% in 2025 MVR 325.61 a month against MVR 342.26. This is arithmetic, not inference: it follows from the two lines above. Source: Derived — mobile revenue ÷ average subscriptions.
OBS The prepaid-to-postpaid migration has stopped Prepaid share fell from 88% in 2016 to 71% in 2022 and has been flat at 72% since. Postpaid is still growing, but so is prepaid. Source: CAM monthly statistics.
OBS Fixed and enterprise is where both operators are growing Fixed, broadband and enterprise revenue rose in almost every quarter for both operators over the period, while mobile revenue was broadly flat. Fixed broadband connections have grown from 16,300 in 2012 to 106,611. Source: Quarterly filings + CAM.
OBS Fixed voice is in terminal decline Fixed telephone lines have fallen from 25,699 in 2012 to 10,992 — a 57% decline, concentrated in but not confined to the Malé region. Source: CAM monthly statistics.
OBS Both operators are profitable and cash-generative Combined 2025 revenue of MVR 5,054m, both reporting EBITDA margins in the high fifties and paying substantial dividends. This is not a market under financial stress. Source: Annual reports, both operators.
Inferred
our reading — could be wrong INF The subscription base is growing on low-value SIMs A base growing 5% while its revenue is flat means the marginal subscription is worth materially less than the average one. Whether that is second SIMs, visitor SIMs, migrant-worker SIMs or genuine price erosion cannot be separated from outside — but all four have the same commercial consequence. Confidence: medium confidence — the arithmetic is firm, the attribution is not.
INF The prepaid base has a visitor cycle in it Prepaid subscriptions rise in December–March and fall in June–July, matching the arrivals season. The market series cannot separate visitor SIMs from resident ones, so this is an association, not a measurement. Confidence: medium-low confidence — consistent pattern, no direct evidence.
INF Price competition is concentrated in short-duration data The published catalogue shows the densest clustering, the deepest promotional flagging and the most inverted per-gigabyte ladders in packs of a week or less. That is where the two operators are visibly fighting. Confidence: medium confidence — inferred from catalogue structure, not from volumes.
INF Value is migrating from connectivity to access moments Both operators sell unlimited-style day passes, app-specific allowances and entertainment bundles priced far above their nominal data cost. That is monetising a behaviour, not a gigabyte. Confidence: medium confidence — clear in product design, unmeasurable in revenue mix.
INF Neither operator has a public retention story No churn figure appears in any filing, annual report or presentation from either operator. That may mean it is not disclosed; it does not follow that it is not managed. But it is the largest hole in the public record of this market. Confidence: high confidence in the absence, no inference about the cause.
Recommended
with the test that would disprove it 01 Treat monetisation, not acquisition, as the binding constraint Adding subscriptions has not added revenue for two years. The commercial plan should be built around revenue per subscription, with the subscription target subordinate to it rather than the other way round. Test: does the next quarter add revenue per subscription, or only subscriptions?
02 Instrument churn before optimising anything else A defined churn event, a monthly cohort survival curve and a value-weighted at-risk population. Without these, retention spend is untargeted and lifetime value is unknowable — and every model on this site that needs churn is running on an assumption. Test: can the business state its monthly churn by value decile? If not, that is the first project.
03 Rebuild the short-duration ladder The per-gigabyte ordering inverts across the published short-pack range, which means some customers are paying more for less and the ladder is not steering anyone upward. Reprice the rungs so that longer and larger is always better value, and let the day passes be priced as access rather than as volume. Test: does moving a customer up a rung raise contribution, or just move revenue between packs?
04 Consolidate the portfolio Multiple packs sit at near-identical price and validity. Each one carries build, billing, support and marketing cost, and adds choice friction. Retire the duplicates, migrate the base, and measure revenue retained rather than assuming it transfers. Test: retire one duplicated pack in a controlled cohort and measure what actually transfers.
05 Run promotions with a control group, always Incrementality cannot be recovered after launch. A matched hold-out group costs a fraction of a percent of take-up and is the difference between knowing a promotion worked and asserting it. Test: what share of last year’s promotions can be shown to have been incremental?
06 Price the visitor and migrant segments deliberately Two million arrivals a year and roughly a third of residents foreign nationals. Both are already served by products — tourist packs and country-specific IDD bundles — but neither is visibly priced as a segment with its own economics. Test: what is revenue per visitor SIM, and how does it compare to a resident SIM over the same period?
07 Keep the growth in fixed and enterprise honest It is the part of both businesses that is actually growing, and it is capital-hungry. Track contribution after the capital, not revenue. Test: return on the fibre and subsea investment, measured against the revenue it carries.
Risks, and what to monitor
What could go wrong
- Price war in short-duration data. The clustering and promotional depth in weekly packs is where a duopoly does most damage to itself. A matched cut takes revenue out of the market permanently.
- Continued ARPU erosion. Two years of flat mobile revenue on a growing base. A third would make the trend structural rather than cyclical.
- Regulatory change to SIM rules. The 2023 SIM registration regulation caps numbers per person. Any tightening removes subscriptions that are already low-value — painless for revenue, visible in every subscription KPI.
- Concentration in a single sector. Enterprise growth is tied to tourism and government. Both are correlated with the same shocks that hit the visitor SIM base.
- Capital intensity of fixed. Fibre, subsea cable and 5G are being funded from a mobile business whose revenue is flat.
The monthly monitor
- Revenue per subscription — the single number this whole programme points at.
- Net adds split by prepaid and postpaid, against the revenue they carry.
- Churn and reactivation by value decile, once instrumented.
- Contribution by product, and the share of revenue in packs under seven days.
- Promotional depth: revenue at list price versus revenue at promotional price.
- Fixed and enterprise revenue against the capital deployed to earn it.
- Prepaid base against arrivals, to size the visitor cycle properly.
What we are not saying. Nothing here is a claim about either operator's
internal management, strategy or performance beyond what their own filings state. We have no
access to their data, and several of the recommendations above may already be in place —
the public record simply does not show it. These are the conclusions the public evidence
supports, framed as what an analyst inside such a business would test first.