Churn war room: how to fight for the subscriber in the MNP era

A single customer’s story

A CEO in Tashkent put it like this: “I was with one operator for ten years. I switched in fifteen minutes through my.gov.” He had no quality complaints. No grievance with support. There was not even a strong offer from the new operator. There was a single conversation with a friend at lunch — the friend said “mine is faster, and porting is free right now”. Ten years of loyalty ended over a lunch break.

This story is not unique. Since mobile number portability went live in Uzbekistan (mnp.uz), and especially after operators started waiving the porting fee in fixed windows — Beeline from 27 December 2025 to 20 February 2026 (beeline.uz), Ucell from 26 March to 30 June 2026 (ucell.uz), Mobiuz from 2 April to 30 June 2026 (mobi.uz) — customer psychology has shifted. The number is no longer an anchor. It is a stand-alone identifier that travels like a name in an address book.

This means the classic retention tools have stopped working. At the moment of decision the customer is no longer loyal. They are comparing. And the decision is taken in minutes, not weeks. The window for a counter-offer is narrow.

What changed

Retention used to be built on the difficulty of switching. The customer had to tell every contact the number had changed. Re-register the number in every app. Explain it to the family. If the customer had a business number tied to advertising, losing it carried reputational cost. These difficulties created switching cost that was enough to retain a substantial part of the base even with a sub-optimal offer.

Now switching cost is essentially zero. The number stays. Contacts do not need to be notified. Business cards do not change. The single point of friction — a visit to the new operator’s office — has been replaced by an online process via my.gov (UzDaily, 21 April 2026). The transfer takes minutes. The financial barrier — operators periodically waive it.

In this reality retention cannot be about friction. Retention now has to be about value and speed of response.

What a war room is

The war room is a borrowing from the crisis management practice in operations companies. In the retention context it is a concrete organisational construct with the following characteristics.

A team of 4-7 people. Retention manager (lead), data analyst, marketer, contact centre representative, product manager, sometimes finance. These people do not work in the war room full time — each has a primary role. But they meet regularly, share a dashboard and take decisions.

A regular routine. Daily standup (15 minutes) — what happened in the last day, which segments are at risk, which actions are scheduled. Weekly review (45 minutes) — conversion by event, retention economics, rule adjustments.

A shared control tower. Not a generic BI report. A specialised dashboard showing the number of customers at each churn risk stage, actions taken in the last day, conversion at each stage, and the economics of retention versus the cost of letting the customer go.

The right to take decisions in the moment. If the retention manager sees that segment X is at high risk and an immediate offer is needed, they should not have to wait for marketing approval. There has to be a pre-approved library of possible actions and an authorisation framework.

Segments where retention work makes sense

Not every customer is worth investing in retention. This is the key point most operators miss.

High value plus high risk of leaving — first priority. These are customers with high ARPU whose behaviour shows signs of dissatisfaction (rising complaints, falling traffic, looking up alternatives on the web). Intensive action is justified.

High value plus low risk — maintenance. Do not run active retention, run loyalty — reward programmes, premium services, white-glove handling at the moments of contact.

Low value plus high risk — economic decision. Often the right answer is to let go. The cost of contact plus the cost of a retention discount exceeds 12-month ARPU. Every telecom operator should have a clear policy on this.

Low value plus low risk — do nothing. This is a healthy segment that needs no special action. Standard service level is enough.

The most common operator error is trying to retain everyone. This dilutes the budget, exhausts the team and produces poor retention economics overall. The discipline of “when to let go” is part of the war room as much as the discipline of “when to retain”.

Signals that work and signals that do not

Not every signal predicts churn equally.

Strong signals (based on similar markets): a sharp 60%+ traffic drop over a week from previously stable behaviour; a third complaint within 30 days without resolution; downgrade to the minimum tariff after a mid-tier; a request for a detailed bill (a classic behavioural signal before leaving); a search for the number portability service.

Weak signals: a single complaint; temporary changes in consumption; a roaming spike (often means a trip, not a departure); one negative dealer interaction.

Useless signals that operators still spend budget on: tenure as an indicator of loyalty (a 10-year customer leaves at the same rate as a 1-year customer); generic ARPU without segmentation (the change matters, not the level); demographic data (age and gender weakly predict churn in B2C telecom).

A good churn model uses 8-15 strong signals and ignores the weak ones. Complex models with 100+ signals look impressive in a demo but in production give marginal improvement over simple models and are harder to maintain.

Speed of response

This is the most important thing. When a customer has decided to leave, they leave fast. The window for a counter-action is hours, not days.

The ideal — a response within 1-4 hours of the first weak signal. This is only possible under two conditions: automation of most decisions and a pre-approved action library. If every retention action requires approval, the window closes.

A realistic target — 24 hours from a strong signal to customer contact, with a pre-approved action and automatic channel selection.

In most operators we observe, response speed is 5-7 days on a strong signal and longer on a weak one. By that point the customer has already decided, and the retention campaign becomes a reactivation campaign with worse economics.

War room architecture

The minimum for launch. An event collector that gathers meaningful events in real time. A trigger engine that applies rules and identifies customers for action. A decision system that picks the right action from the pre-approved library. A channel orchestrator that delivers in the correct channel (push, SMS, call, dealer ping). An audit log that records what happened. A performance dashboard for the war room team.

This is not buying one system. It is an assembly of components, many of which already exist at any operator. Often the new component is only the trigger engine with rules and the dashboard. The rest is delivered through integration.

The cost of this minimum is 8-12 months of work and a few FTE for the build. Payback is a measurable reduction in churn of 10-30% in the pilot segment within 6 months of the war room routine starting.

When the war room is not needed

If overall churn rate is below 1% per month and stable, the war room operating cost may not justify the investment. Here it is more correct to work on loyalty.

If the operator’s ARPU is low (below 30,000 UZS per month), retention economics do not work even with a perfect retention machine.

If master data on the customer does not reconcile between systems, retention actions will keep making mistakes — calling customers who have already left, sending offers to those who already decided to stay. This destroys the team’s trust in the system.

If the team has no head of retention with a real mandate and P&L, the war room exists on paper, not in practice.

If the market has no MNP or MNP is inactive (transfer volumes are small), the urgency is missing. Operators invest in retention proportionally to the intensity of porting.

Where to start this quarter

If the operator decides the war room is needed, the sequence is simple.

The first 30 days. Signal audit: which events can we collect in real time without additional investment. A list of 5-10 strong triggers. A map of existing response channels. Appointment of a retention manager with authority.

The next 30 days. Pilot on one high-value segment (for example, postpaid customers with ARPU above the median). A trigger engine with 3-5 rules. A pre-approved action library. Daily review.

The third month. Measurement. Conversion at each stage. Cost of retention. Saving from retained customers. Data quality. Readiness to scale.

By the end of the quarter there is an answer: does the war room work at this operator and does it make sense to scale to other segments. If it does — scaling over 2-3 quarters. If not — a diagnostic on what did not work (data, speed, owner, economics) and a decision on an alternative approach.

How SamaraliSoft can help

Subscriber Intelligence & Churn War Room — a concrete 90-day programme. An audit of current retention, selection of a pilot segment, building of a minimal viable war room, daily routine, weekly review with the financial lens, and a scaling decision. Not “we will build you a system” but “we will set up an operating routine that lowers churn in the pilot within a quarter”.

Related reading

• /en/insights/telecom-subscriber-intelligence-operating-model/ — operating model around data
• /en/cases/telecom-subscriber-360/ — a real Subscriber 360 case for retention
• /en/insights/telecom-growth-after-connectivity/ — where the money sits in telecom beyond connectivity
• /en/insights/telecom-retention-economics/ — when retaining is more expensive than letting go

Sources

• MNP.UZ — mobile number portability
• Beeline UZ — porting conditions, 27 December 2025
• Ucell — porting conditions from 26 March 2026
• Mobiuz — MNP price and conditions
• UzDaily — online mobile number transfer via my.gov