The models a commercial team would actually run.
Six engines, each with its assumptions on the surface rather than buried in it. Where the public record supports the model it is anchored on observed data; where it does not, the model is labelled as a framework and the missing dataset is named.
01 — Revenue decomposition
anchored on observed dataStarting from the market's 2025 position: 808,699 average mobile subscriptions at MVR 325.61 of mobile service revenue per subscription per month. Move the two growth rates and the revenue path follows.
| Revenue change over 5 years | ARPU -8% | ARPU -6% | ARPU -4% | ARPU -2% | ARPU 0% | ARPU +2% |
|---|---|---|---|---|---|---|
| Subs 0% | −34% | −27% | −18% | −10% | +0% | +10% |
| Subs +2% | −27% | −19% | −10% | −0% | +10% | +22% |
| Subs +4% | −20% | −11% | −1% | +10% | +22% | +34% |
| Subs +6% | −12% | −2% | +9% | +21% | +34% | +48% |
| Subs +8% | −3% | +8% | +20% | +33% | +47% | +62% |
Base position derived from Ooredoo Maldives and Dhiraagu mobile service revenue (quarterly filings, CMDA InformInvestor) over average monthly mobile subscriptions (Communications Authority of Maldives). The denominator counts SIMs, so this is revenue per active SIM across both operators — not either operator's own reported ARPU, and not revenue per person.
02 — Pricing and elasticity
stated demand modelQ = Q₀ · (P/P₀)ε. Elasticity is an input here, not an estimate.
With constant elasticity, revenue has no interior maximum — if |ε| > 1 revenue always rises as price falls, and if |ε| < 1 it always rises as price rises. That is a property of the functional form, not of the market, and it is exactly why the contribution optimum (which does exist, because variable cost bites) is the one a commercial team should be looking at.
03 — Customer lifetime value
framework with observed anchorsARPU is pre-set to the market's 2025 revenue per mobile subscription and margin to the reported EBITDA margin, so the defaults are grounded. Churn, acquisition cost and retention spend are not published by either operator and are the reader's assumptions.
The curve is steep at low churn and flat at high churn, which is the practical point: at 6% monthly churn an operator is replacing its base every sixteen months and improving ARPU barely moves lifetime value, while at 2% the same ARPU improvement compounds over four years. Where an operator sits on that curve decides whether the commercial priority is retention or monetisation — and neither Maldivian operator publishes the churn number that would locate them on it.
04 — Churn and retention targeting
simulated baseA synthetic base of 4,000 prepaid customers, scored by an interpretable logistic model. Choose which customers a fixed retention budget should buy.
| Customer | Spend / month | Recharges / month | Days since recharge | Tenure | Data trend | Churn probability | Value at risk |
|---|---|---|---|---|---|---|---|
| C-2365 | MVR 486 | 1.0 | 45 | 11m | −65% | 80% | MVR 2,653 |
| C-0765 | MVR 419 | 0.6 | 52 | 25m | −65% | 85% | MVR 2,442 |
| C-0264 | MVR 406 | 0.8 | 51 | 13m | −59% | 85% | MVR 2,362 |
| C-0862 | MVR 423 | 0.5 | 54 | 37m | −52% | 80% | MVR 2,321 |
| C-1281 | MVR 416 | 1.0 | 51 | 4m | −31% | 77% | MVR 2,201 |
| C-2134 | MVR 389 | 1.2 | 48 | 8m | −59% | 83% | MVR 2,197 |
| C-1857 | MVR 437 | 2.0 | 51 | 42m | −65% | 71% | MVR 2,116 |
| C-2382 | MVR 407 | 0.5 | 47 | 21m | −41% | 75% | MVR 2,087 |
| C-1880 | MVR 420 | 0.8 | 58 | 43m | −30% | 72% | MVR 2,073 |
| C-1142 | MVR 427 | 0.7 | 53 | 50m | −43% | 70% | MVR 2,044 |
Scoring model: logit(p) = -1.15 + 0.055·days-since-recharge − 0.34·recharges-per-month − 0.021·data-trend − 0.018·tenure − 0.0016·spend. Coefficients are stated, not fitted. Switching the policy from value at risk to churn probability raises the average churn probability of the contacted group and lowers the value protected — which is the result the panel exists to show.
05 — Promotion incrementality and cannibalisation
worked exampleTwo existing products and one promotional price between them. The question is not how many customers took the promotion — it is what they were worth before they took it.
| Revenue bridge | Customers | Was worth | Now worth | Effect |
|---|---|---|---|---|
| Switched from product A at MVR 300 | 9,000 | MVR 2.7m | MVR 3.15m | +MVR 0.45m |
| Switched from product B at MVR 400 | 6,000 | MVR 2.4m | MVR 2.1m | −MVR 0.3m |
| New to the category | 5,000 | — | MVR 1.75m | +MVR 1.75m |
| Incremental revenue | 20,000 | MVR 5.1m | MVR 7m | +MVR 1.9m |
Set the promotion below product A's price with most take-up coming from A's existing customers and the incremental line turns negative while the headline stays large. That is the case worth being able to spot before launch, not after: a campaign can grow subscriptions, be reported as a success, and reduce revenue. The one input that decides it — what switchers were spending before — is knowable only from the operator's own pre-period data, which is why a promotion needs a defined control group at design time.
06 — Commercial opportunity ranking
analyst ratings, your weightsNine initiatives, five criteria, weights you set. There is no objectively correct weighting — the value of the exercise is that changing the weights changes the answer visibly, and forces the trade-off to be argued rather than assumed.
| Initiative | Size of the prize | Underlying growth | Margin quality | Feasibility | Competitive headroom | Score |
|---|---|---|---|---|---|---|
| 01Prepaid-to-postpaid migration | 4 | 4 | 5 | 4 | 3 | 82 |
| 02Retention and reactivation programme | 4 | 3 | 5 | 3 | 4 | 77 |
| 03ARPU defence on the core data pack | 5 | 2 | 5 | 4 | 2 | 75 |
Stop the erosion in revenue per subscription by re-laddering the mid-tier data packs rather than adding volume to them. Evidence: Mobile revenue was flat in 2025 (+0.1%) while subscriptions grew 5.2%: revenue per subscription fell 4.9%. | ||||||
| 04Visitor and short-stay monetisation | 3 | 4 | 4 | 4 | 3 | 73 |
| 05Portfolio simplification | 3 | 1 | 4 | 5 | 4 | 68 |
| 06Migrant-worker segment | 4 | 3 | 3 | 4 | 3 | 68 |
| 07Entertainment and gaming bundles | 3 | 4 | 3 | 4 | 3 | 68 |
| 08Fixed broadband expansion | 4 | 3 | 4 | 3 | 2 | 66 |
| 09Enterprise and resort connectivity | 4 | 3 | 4 | 2 | 2 | 62 |
Ratings are on a 1–5 scale and are analyst judgements, not measurements; the evidence line under each initiative says what the judgement rests on. Score = Σ(rating/5 × weight) ÷ Σweights, expressed out of 100. Under any weighting that puts meaningful value on margin quality, ARPU defence and retention rank at or near the top — which is the finding, not the formula.
Customer segmentation — the framework, not a finding
| Dimension | Definition | Source system |
|---|---|---|
| Recency | Days since the last revenue-generating event | Recharge and usage records |
| Frequency | Recharges and pack purchases per month | Recharge records |
| Monetary | Rolling three-month spend | Billing records |
| Engagement | Data volume, voice minutes, app mix, trend in each | Network usage records (CDR/xDR) |
| Tenure | Months since activation | Subscriber master |
| Portfolio | Which products held, and how that has changed | Product subscription records |
Implementation: score each subscriber on each dimension over a rolling window, cluster or rule-band the scores, and re-run monthly so a customer can move between segments — a segmentation that never moves is a label, not a behaviour.
| Segment | Spend | Behavioural signature | Commercial action |
|---|---|---|---|
| High value | Top decile of monthly spend | Frequent recharges, high data, long tenure | Protect. Proactive service, early access, no discount needed. |
| Growth | Mid spend, rising | Data use trending up, buying larger packs over time | Upsell the next rung of the ladder before they find it themselves. |
| Price sensitive | Low to mid | Concentrated on promotional packs, switches on price | Serve with targeted offers; never with an across-the-board price cut. |
| Heavy data | Mid to high | Consistently exhausts allowances, buys boosters | Move onto a larger recurring plan — cheaper for them, steadier for the operator. |
| Entertainment & gaming | Mid | Skewed to streaming, social and gaming allowances | Bundle content; this is where willingness to pay is least commoditised. |
| At risk | Any | Recharge frequency falling, days since last recharge rising | Retention offer, sized by value at risk rather than by churn probability. |
| Dormant | None currently | Previously active, no revenue event in the last cycle | Reactivation, and a decision on when to stop paying to keep the SIM alive. |
| Low value | Bottom decile | Minimal recharge, minimal usage, often a second SIM | Serve digitally at near-zero cost. Do not spend retention budget here. |
A segmentation earns its cost only if a different action follows from each segment. Two segments with the same action should be one segment.
Product lifecycle — the gates a pack should pass
The published catalogue shows several packs that a working version of this process would have caught: near-identical price and validity to a neighbouring pack, and per-gigabyte ladders that invert as duration rises. Portfolio simplification is on the opportunity list above for that reason.