Executive Summary
This analysis examines 12 months of customer behaviour across 7,032 telecom accounts.
At 26.6% churn, the business is losing
$139,131 in monthly recurring revenue
from departed customers. Three structural drivers account for the majority of attrition:
contract type, payment method, and the critical first-year window.
Targeted interventions on these levers — without blanket discounting — offer the highest-
probability path to measurable churn reduction.
Finding 1 — Contract Type is the Dominant Churn Driver
Month-to-month contracts churn at a rate nearly 8× that of two-year contracts.
This is the single most predictive variable in the dataset.
Implication for strategy: Month-to-month flexibility is a churn engine.
The most impactful commercial intervention is an incentive programme that converts
month-to-month subscribers to 12- or 24-month terms — not retention discounts applied
uniformly after churn intent has already been signalled. Churn prevention is cheaper
than win-back; win-back is cheaper than acquisition.
Finding 2 — The First 12 Months are the Danger Window
Customers in their first year churn at nearly 3× the rate of those
who have stayed four years or more. If a customer survives year one, the business
has likely earned their long-term loyalty.
Implication for strategy: The first 90 days are the highest-leverage
onboarding window. Proactive check-ins, service usage nudges, and a structured 6-month
review call — none of which require discounting — can materially shift first-year
survival rates. The data does not reveal why early customers churn; that
requires a qualitative overlay (NPS / CSAT). But the when is unambiguous.
Finding 3 — Payment Method Signals Churn Risk
Electronic check users churn at 45% — more than double the rate of
automatic payment customers. This is partly a proxy for contract type (month-to-month
customers disproportionately use electronic check), but the signal holds even after
controlling for contract.
Implication for strategy: Nudging electronic check users toward
automatic bank transfer or credit card payment serves two goals simultaneously:
it reduces payment failure risk and is correlated with lower churn. A simple
incentive (e.g., one month fee credit for switching to auto-pay) is a low-cost
intervention with measurable ROI.
Finding 4 — Higher Spend Does Not Mean Lower Churn
Premium-tier customers ($70–$95+/month) churn at rates comparable to
or exceeding mid-tier customers, and they represent a disproportionate share of
monthly revenue at risk.
Implication for strategy: High-value customers are not self-retaining.
A premium account with Fiber Optic + multiple add-ons on a month-to-month contract
is a high-risk combination. Targeted account management — not self-service — is
appropriate for this segment. The revenue-at-risk line on the chart makes this
concrete: premium-tier churn is a larger absolute problem than the rate alone suggests.
Highest-Risk Segments
Cross-tabulation of contract type and payment method reveals the specific
combinations requiring immediate intervention.
| Contract | Payment Method | Customers | Churn Rate |
|---|
| Month-to-month | Electronic check | 1,850 | 54% |
| Month-to-month | Bank transfer (automatic) | 589 | 34% |
| Month-to-month | Credit card (automatic) | 543 | 33% |
Month-to-month + Electronic check is the canonical high-risk customer: flexible contract,
manual payment, typically short tenure. This cohort should be the first target of any
retention programme.
Recommended Next Steps
- Conversion campaign — Month-to-month → Annual contract.
Model the incentive cost against the lifetime value delta of 12- vs 24-month
contracts before setting the offer value.
- Auto-pay migration drive.
Single-touch digital nudge to electronic check users. Measure adoption rate
and 90-day churn delta in the converted cohort.
- 90-day onboarding programme.
Structured check-in sequence for new customers. Instrument it so you can
measure first-year survival improvement vs the 47% baseline churn in that cohort.
- Premium segment account management.
Flag Fiber Optic + month-to-month + high spend combinations for proactive
outreach — do not leave them in a self-service queue.
- Qualitative overlay.
This analysis tells you who churns and when.
It does not tell you why. A structured exit survey or NPS analysis
of churned accounts is the necessary next step to validate the intervention hypotheses.
Data & Methodology Notes
Source: IBM Watson Analytics Telco Customer Churn dataset, 7,043 records.
11 records with tenure=0 and blank TotalCharges were excluded (no billing history,
none churned; <0.2% of the dataset). SeniorCitizen encoding normalised from
integer to categorical. All figures are descriptive; churn rates are proportions
from historical data, not predictions. The dataset does not include cohort start
dates, so time-series trending is not possible with this extract alone.
Churn prediction modelling (logistic regression, random forest) is available
as a Phase 2 deliverable.