Micro-segments instead of the 'mass base': what granularity wins

What “segmentation” usually means today

Look at segmentation at a mid-to-large regional telecom and it is usually built across 5-10 top-level groups. Postpaid premium, postpaid mass, prepaid active, prepaid low, business SME, business corporate, youth, family. These 5-10 cover the whole base.

Each group has a shared tariff portfolio, shared marketing campaigns, shared retention offers. This lets the marketing team operate at scale — one campaign per segment, a clear target, a visible result.

The problem is that each such group actually contains very different customers. Postpaid mass is the young professional in Tashkent with active digital behaviour, the pensioner in a district centre, the student. Their needs, usage patterns and reactions to offers are different. A universal campaign on this segment ignores half the audience and works only for an averaged customer who is rare in reality.

This is the efficiency boundary that 5-10 segments hit.

What changes at 50-100 segments

If the same base is split into 50-100 segments, each becomes specific. Not “postpaid mass” but “active digital user 25-35 in Tashkent with roaming spend” and separately “family user 35-45 in Samarkand with international calls to Russia”. Different segments with different approaches.

What is gained:

Communication relevance. Each offer goes to the segment for which it is meaningful. Open rate up, conversion up, campaign ROI up.

Tariff product fit. A tariff optimal for one segment is not optimal for another. With 50-100 segments a tariff portfolio with real fit per segment is buildable, instead of one compromise mass product.

Churn cause identification. Different segments churn for different reasons. With 5-10 these reasons mix and become unreadable. With 50-100 they are visible and addressable.

Pricing. Different segments have different price elasticity. 50-100 segments allow individual optimisation.

What blocks the move to 50-100

Several reasons most operators stay at 5-10.

Operational complexity. Managing 50 segments looks unmanageable — 50 different campaigns, 50 different offers. Fear of complexity.

Data quality. Many segmentation attributes (behaviour, usage, geographic pattern) need data quality that often is not there.

The marketing team does not scale. If every campaign requires manual work, 50 campaigns mean a 50-fold load.

Vendor tools do not support it. Many campaign management tools are optimised for 5-10 segments with large audiences. Managing 50-100 micro-segments requires a different architecture.

The board thinks in big numbers. “A segment of 200 customers” is read as insignificant and the investment is rejected. Even though 100 such segments is the entire base.

What the move requires

Management complexity is the real question. Real scaling to 50-100 segments needs operations automation. Not “the marketer designs 50 campaigns”, but a system that automatically assigns customers to segments and applies segment-specific rules.

Which means:

Segmentation as a continuous process. Each customer is automatically reassessed weekly or monthly based on behaviour and lands in the right segment. Not manually once a year.

A modular tariff portfolio. Not 50 separate tariffs but 5-7 base components combined per segment. Simpler product management.

Campaign automation. Most communication is trigger-based, not calendar-based. Triggers are tied to segments and events.

Tooling that supports many segments. Customer Data Platform or equivalent. Without the right tools 50-100 segments stays a theory.

A team with a different mindset. Not a “mass marketer”, but a data-driven product manager working with micro-segments as products.

A realistic path to granularity

Not “50 segments tomorrow”. Phased.

Months 1-6. A vertical break of one existing group. For example, postpaid mass splits into 8-10 sub-segments. A pilot — without infrastructure for all 50, start with one group.

Months 7-12. Building infrastructure. Customer Data Platform, automation, segmentation engine. A vertical investment.

Months 13-18. Expansion to other groups. Each main group splits into 5-10 sub-segments. That gives 50-80 micro-segments in total.

Months 19-24. Optimisation. Tracking which micro-segments perform and which do not. Merging the indistinguishable, splitting the too coarse.

By two years in the operator has working segmentation at 50-100 micro-segments. Marketing efficiency is meaningfully higher, retention addresses real causes, the tariff portfolio fits better.

Where this often breaks

A break into segments without action plans. You have 50 segments but 40 of them have no separate strategy. Theatre — segments without actions.

Splitting to illusory granularity. If a segment has 200 customers but is statistically indistinguishable from a neighbour at 250, the split is fake.

Tooling without operating discipline. CDP bought, but no process for marketing across 50 segments. A tool without a process does not help.

Segments built without business angles. The data scientist builds segmentation from clean data without understanding where the business value lives. Academically interesting, commercially useless.

Cross-segment overlap. The customer falls into 5 segments at once (different rules) and gets 5 different communications. Chaos rather than personalisation.

When not to launch a segmentation reform

If the data foundation is weak, segmentation is on flawed data and actions are wrong. Foundation first.

If the marketing team has no capacity for the new mindset, the reform hits the human factor. Training or recruitment is needed.

If C-level is not ready for an 18-24 month investment without visible ROI in the first 6 months, the reform dies in the pilot.

If the competitive environment requires aggressive quick wins, the reform on a longer horizon is a luxury.

If a CDP or equivalent does not exist and a vendor rollout will take 9-12 months, the reform on an 18-month horizon becomes unrealistic.

Discussion points for the committee

How many segments are in active use today and how distinguishable are they really? If hard to answer — diagnostic.

Which 3 segments have the worst performance in current campaigns? They are split candidates.

Is the organisation ready for a 12-18 month trajectory with measurable result not in the first quarter? Without that commitment the reform does not start.

Who owns segmentation as a function — data team, marketing, product? Without an owner nothing happens.

What is the data foundation — is there a master customer ID, behavioural events, attributes for segmentation? If not — that is the first thing.

How SamaraliSoft can help

Micro-Segmentation Strategy & Operating Model — analysis of existing segmentation (what is there, where the gaps are), design of a micro-segmentation framework against the operator’s structure, evaluation of tooling (build / buy / integrate) for CDP and automation, organisational design of the new marketing operating model, and a phased rollout with a pilot on one group over 90-180 days.

Related reading

• /en/insights/telecom-subscriber-intelligence-operating-model/ — operating model
• /en/insights/telecom-regional-segmentation/ — regional segmentation
• /en/insights/telecom-nba/ — NBA as context
• /en/insights/telecom-offer-fatigue/ — why mass marketing does not work

Sources

• Digital Government Portal — digital infrastructure UZ
• UzDaily — Central Asia telecom market
• Ministry of Digital Technologies — telecom in UZ
• MNP.UZ — number portability