Deep Dive · October 2026
How to Measure Hosting Churn and Track Customer Migration
Learn how hosting companies can combine account churn metrics with domain attrition, observed website migration and market-level provider movements.
Target audience
Product & Commercial Leaders
Publication
October 2026
Author
Nadya Frost
Executive Summary
Billing systems only tell you when an invoice run fails or a contract ends. Not when a customer stopped using your services and is about to leave. Separating structural churn from actionable migrations helps formulate the right interventions:
- Low churn rates is not safety: In Europe, 18 of 50 major providers posted attrition rates of >1% per quarter, but are still net-donors to competitors. → Understand who you're losing to and why.
- The same migration can yield very different capture: Comparing two major brands' inflow, one saw 80% domains switch over their mail, and the other only 20%. → Invest in onboarding automation and bundling to lock in revenue.
- Services may move before the domain transfers: Weeks before a DNS change, up to 14% of mail services and 4% of websites are already running on new infrastructure. → Treat infra changes as a warning signal, and build winback flows for key products.
Based on 13 weeks of observations to 19 October 2026: the top 50 European providers (32.7M domains) and migration cohorts at two major operators.
Definitions
How to measure churn?
While a healthy intake of new customers is key for running a sustainable business, losing too many too early creates a vicious circle: CPAs go up, LTVs trend down, requiring brands to acquire ever more customers. We've already explored how ShareShift insights help brands acquire more customers; this article explores ways of growing by reducing attrition, and taking market share from competitors.
The next question after "why focus on churn?" is, how to measure it. And here much room for interpretation exists, risking misleading churn statistics and conclusions.
Here is a practical 5-point measurement system that brings together different views on churn:
Commercial Takeaway
Stop expecting any single measure to answer every churn-related question: who, what, when, why. Instead, combine internal billing data with behavioural signals and observed infrastructure changes.
5-Point Churn Measurement System
- Observed domain changes: Domain attrition, incoming and outgoing migrations, and technical changes in provider attribution.
- Service signals: Changes in web and mail configuration, interpreted according to stated detection rules.
- Market context: Competitor movements, market-level portfolio growth and gross migration intensity.
- Account outcomes: Cancellations, renewals, account churn and revenue churn sourced directly from billing and CRM.
- Customer explanations: Documented cancellation reasons, survey responses and win/loss feedback.
In this article, we unpack the first three measurements: provider- and market-level observed domain attrition trends, as well as migration-based growth based on Q3'26 data. The last two analyses need to be conducted internally. See the full list of churn-related metrics and sources under Methodology, at the end of this article.
Observation Architecture
How to track provider migrations?
ShareShift observes domains and their technical configuration over time. The result is a clear from-to tracker that highlights trend changes over time. Here's how it is built:
A: Establish a baseline
Attribute all domains in your universe to service providers; define attribution rules, exclusions and observation window.
B: Run observations over time
Identify domains whose provider attribution changes between scheduled observation intervals.
C: Examine related service signals
Track changes in other configuration around the migration event; e.g., if you track websites, check what's happening to email.
D: Attribute the movement
Identify origin and destination providers for each service, applying confidence rules and excluding ambiguous cases like CDNs.
E: Aggregate the evidence
Calculate provider- and market-level attrition, incoming and outgoing migration, net migration and gross migration intensity.
Benchmark Analysis
Growing by losing less and taking share
An operator's churn volume consists of natural attrition (structural churn) and its trading balance (net effect of migrations).
To make sense of where providers actually gain and lose ground, we have analyzed domain attrition and migration-based growth for top 50 European vendors. Jointly, these vendors aggregate a portfolio of 32.7 million domains.
The graph below maps them on a canvas, followed by a simple 2x2 matrix that explains observed patterns.
- The X axis maps the net-migration effect as a percentage change in the provider's active base during the period. Zero is net-neutral growth.
- The Y axis is the attrition rate, with the median of 0.76% per quarter.
Data table
| Name | Attrition Pct | Bubble size | Net Migration Pct |
|---|---|---|---|
| Provider 1 | 0.2 | 1,057,706 | 0.06 |
| Provider 2 | 0.23 | 1,141,288 | 0.69 |
| Provider 3 | 0.24 | 498,011 | 0.82 |
| Provider 4 | 0.27 | 545,534 | 0 |
| Provider 5 | 0.29 | 2,778,357 | -0.02 |
| Provider 6 | 0.31 | 340,037 | 0.02 |
| Provider 7 | 0.39 | 251,382 | -0.06 |
| Provider 8 | 0.4 | 370,200 | 0.69 |
| Provider 9 | 0.42 | 126,422 | -0.1 |
| Provider 10 | 0.46 | 213,627 | -0.33 |
| Provider 11 | 0.49 | 454,390 | 0.06 |
| Provider 12 | 0.51 | 331,315 | 0.04 |
| Provider 13 | 0.55 | 353,660 | -0.26 |
| Provider 14 | 0.55 | 171,081 | -0.19 |
| Provider 15 | 0.56 | 172,288 | 1.25 |
| Provider 16 | 0.62 | 404,826 | 2.74 |
| Provider 17 | 0.63 | 512,813 | -0.21 |
| Provider 18 | 0.64 | 1,132,333 | -0.31 |
| Provider 19 | 0.67 | 135,337 | -0.4 |
| Provider 20 | 0.69 | 128,279 | -0.11 |
| Provider 21 | 0.7 | 291,310 | -0.45 |
| Provider 22 | 0.71 | 432,243 | -0.32 |
| Provider 23 | 0.72 | 368,766 | -0.45 |
| Provider 24 | 0.74 | 4,489,074 | -0.18 |
| Provider 25 | 0.76 | 247,809 | 0.37 |
| Provider 26 | 0.76 | 131,245 | 0.14 |
| Provider 27 | 0.77 | 214,827 | -0.36 |
| Provider 28 | 0.79 | 557,927 | -0.13 |
| Provider 29 | 0.84 | 119,724 | -0.69 |
| Provider 30 | 0.88 | 163,517 | -0.18 |
| Provider 31 | 0.95 | 187,342 | 0.56 |
| Provider 32 | 0.96 | 782,567 | 0.18 |
| Provider 33 | 0.97 | 116,575 | 0.27 |
| Provider 34 | 1.01 | 263,869 | -0.58 |
| Provider 35 | 1.02 | 205,578 | 5.75 |
| Provider 36 | 1.06 | 251,104 | -0.91 |
| Provider 37 | 1.11 | 312,263 | -0.19 |
| Provider 38 | 1.25 | 831,700 | 0.63 |
| Provider 39 | 1.25 | 165,743 | -1.08 |
| Provider 40 | 1.54 | 165,684 | -1.23 |
| Provider 41 | 1.78 | 293,460 | -0.49 |
| Provider 42 | 1.8 | 170,529 | -0.76 |
| Provider 43 | 2.06 | 808,355 | -0.99 |
| Provider 44 | 2.37 | 144,008 | -0.15 |
| Provider 45 | 2.48 | 4,617,558 | -2.25 |
| Provider 46 | 3.02 | 495,155 | -2.73 |
| Provider 47 | 3.75 | 386,509 | -3.27 |
| Provider 48 | 6.38 | 915,290 | 2.55 |
| Provider 49 | 8.02 | 630,759 | -4.59 |
| Provider 50 | 15.14 | 164,586 | 2.41 |
Source: 13-week actuals anonymized; weeks Jul'27 - Oct'19 2026. Exclusions: marketplaces, MSPs and infrastructure players; all Cloudflare-related DNS changes.
Strategic Archetypes
Which quadrant are you trading in?
A hosting company can have an exceptionally loyal base and still bleed market share through net-negative migration; another can pull in thousands of transfers with aggressive promos, only to watch them cycle out just as fast. To better understand the dynamics of product stickiness vs competitive trading across European brands, we introduce a simple matrix that hinges on on two clear thresholds:
- Trading balance (0% net migration): Are you winning more domains from rivals than you lose, or are you a net donor?
- Baseline retention (0.765% sample median): Are you dropping domains at an above-average rate?
Knowing your quadrant tells your commercial team what problem you actually need to solve: whether you are dealing with a leaky product, or simply failing to compete when customers look for a new home.
22% of Providers
The Best Performers
The benchmark quadrant. These brands combine solid product stickiness with aggressive transfer flows, reliably taking market share.
Hypothesis to test: Multi-service activation. Portfolios pairing domain registration with active mail routing and simple site creation build structural moats that deter competitor displacement.
28% of Providers
The Slow Decline
These operators enjoy an exceptionally sticky, loyal customer base that rarely cancels outright. Yet when domains move between providers, they are net donors to rivals.
Hypothesis to test: Inbound proposition fatigue. A reliable billing base breeds complacency; without competitive introductory offers and high-converting onboarding funnels, established brands steadily lose ground.
14% of Providers
The Waystations
These providers win substantial incoming volume, but cannot retain it. Domains cycle through, yielding high gross movement with low long-term portfolio stability.
Hypothesis to test: Promotional offer arbitrage. Aggressive loss-leader campaigns attract price-sensitive customers; without sticky service activation, domains churn out the moment introductory pricing resets.
36% of Providers
The Fast Decline
The most vulnerable group. More than a third of the observed providers suffers from high deactivations paired with consistent net negative migration.
Hypothesis to test: Pricing shocks and consolidation friction. Spikes in dual-metric decay frequently follow aggressive renewal price increases, legacy platform migrations, or post-acquisition support degradation.
Curious which quadrant your portfolio occupies this quarter?
Contact the ShareShift team to benchmark your brand against the market and review your migration corridors.
Service Attach Relative to DNS Migration
What happens to web and email?
A hosting customer is worth substantially more than a domain renewal; tracking what happens to mail, web, and auxiliary services around a transfer is critical for predicting customer lifetime value.
To understand that sequence, ShareShift analyzed DNS migration trends as a stand-in measure for inter-registrar domain transfers. The study covers 40,000+ domain migration events of two major brands across February - October 2026. Each sample covers visible mail and web service attribution from five weeks before to five weeks after the DNS switch.
Actionable learnings:
- DNS switching triggers immediate service capture.
At week zero, old-provider mail and web collapse below 2% across both brands. For Brand B, mail on the new provider jumps from 14% pre-migration to 80% afterwards. Even third-party email takes a hit: Brand B compresses "other provider" mail from 12.5% pre-migration down to 8% at week zero.
→ Done right, post-migration onboarding can pull external revenue directly into the host ecosystem. - Automated provisioning (likely) explains the divergence.
Brand B likely automates mailbox provisioning and MX record updates at DNS switch: its new-provider mail jumps by 66 percentage points in week zero, whereas Brand A rises by just 14 percentage points.
→ Default bundling secures initial uptake, but internal data is required to prove active mailbox utilization and downstream churn reduction. - Service leakage post-migration is real, but uneven.
Active mail footprint for Brand A drops sharply: from 76% pre-migration to 59% at week zero and up to 64% by week +5, leaving inactive mail up nearly 12 percentage points. For Brand B, web hosting sees post-migration drop-off: new-provider web peaks at 33% at week zero and drifts down to 26% by week +5, with total inactive web rising from 47% to 59%.
→ Re-activating lapsed services that customers ran before transferring represents direct ARPU expansion. - Pre-migration service adoption signals intent.
A measurable slice of customers points services to the destination host weeks before moving DNS. At week -5, Brand B already hosts 14% of incoming mail and 4% of incoming web; Brand A already hosts 6% of mail and 3% of web.
→ Seamless staging environments and onboarding flows capture these high-intent workloads early.
Data table
| Name | New Provider | Old Provider | Other Provider | Unknown Provider |
|---|---|---|---|---|
| W-5 | 6.15% | 17.26% | 21.84% | 30.36% |
| W-4 | 5.87% | 17.18% | 22.07% | 26.45% |
| W-2 | 5.58% | 16.83% | 21.47% | 29.47% |
| W0 | 19.78% | 1.7% | 22.83% | 14.73% |
| W+1 | 20.86% | 1.78% | 24.71% | 13.98% |
| W+2 | 21.5% | 1.8% | 24.83% | 14.63% |
| W+3 | 22.14% | 1.89% | 25.63% | 14.23% |
| W+4 | 22.18% | 1.87% | 25.27% | 14.58% |
| W+5 | 22.21% | 1.8% | 25.32% | 14.79% |
Data table
| Name | New Provider | Old Provider | Other Provider | Unknown Provider |
|---|---|---|---|---|
| W-5 | 2.97% | 11.13% | 11.54% | 23.97% |
| W-4 | 2.74% | 11.28% | 10.14% | 25.03% |
| W-2 | 2.5% | 11.12% | 10.07% | 23.8% |
| W0 | 17% | 1.21% | 12.37% | 14.28% |
| W+1 | 15.58% | 1.52% | 12.97% | 13.96% |
| W+2 | 15.17% | 1.54% | 13.61% | 14.25% |
| W+3 | 15.51% | 1.51% | 14.05% | 14.21% |
| W+4 | 15.28% | 1.43% | 14.11% | 14.56% |
| W+5 | 15.61% | 1.37% | 14.28% | 14.58% |
Data table
| Name | New Provider | Old Provider | Other Provider | Unknown Provider |
|---|---|---|---|---|
| W-5 | 13.94% | 20.63% | 12.46% | 37.3% |
| W-4 | 13.78% | 20.65% | 12.5% | 37.08% |
| W-2 | 13.07% | 20.93% | 12.35% | 36.31% |
| W0 | 80.13% | 0.6% | 7.89% | 7.41% |
| W+1 | 79% | 0.77% | 9% | 8.18% |
| W+2 | 78.26% | 0.83% | 9.38% | 8.38% |
| W+3 | 77.92% | 0.85% | 9.55% | 8.5% |
| W+4 | 77.92% | 0.88% | 9.69% | 8.73% |
| W+5 | 77.72% | 0.86% | 9.75% | 8.81% |
Data table
| Name | New Provider | Old Provider | Other Provider | Unknown Provider |
|---|---|---|---|---|
| W-5 | 4.29% | 12.56% | 8.9% | 26.92% |
| W-4 | 4.21% | 12.53% | 8.89% | 26.65% |
| W-2 | 3.93% | 12.09% | 8.58% | 25.85% |
| W0 | 33.19% | 0.92% | 5.93% | 6.19% |
| W+1 | 29.31% | 1.27% | 6.71% | 6.55% |
| W+2 | 27.03% | 1.37% | 7.11% | 6.55% |
| W+3 | 26.01% | 1.44% | 7.42% | 6.43% |
| W+4 | 25.84% | 1.47% | 7.38% | 6.56% |
| W+5 | 25.71% | 1.53% | 7.38% | 6.56% |
In this benchmark, Brands A and B are two major European operators, anonymized. The old provider is the origin host, the new provider is the destination, and Other represents third parties (e.g., Microsoft 365, Google Workspace, Shopify). Unknown is the long tail of self-hosted services and unmapped vendor infrastructure. Parked and placeholder pages are not excluded, and may inflate total web attach rates before/after the domain migration.
European Market Data
How market growth shapes provider-level churn
Provider-level metrics do not tell the full story. A 12% annual attrition rate (3% per quarter) can signal either elevated churn or exceptional retention, depending on how stable the overall market is. Some established European ccTLDs consistently show high renewal rates and rare provider switches; others are showing considerably more movement.
To separate genuine operational performance from local macro currents, registrars and hosts must benchmark their own performance against market baselines. Our data shows how much European ccTLDs diverge across attrition rates: from 2% per quarter in .at to 7.1% in .pl.
These are observed domain deletions relative to the active registry base, as tracked by ShareShift. Registrars and hosting providers can compare their own attrition rate (excl. migrations) to the benchmark, and ask whether a gap reflects their own performance, or a macro trend.
Combined with market expansion rate and gross migration volumes, market-wide attrition rates can show Which European Markets Are Growing, and Which Are Just Reshuffling?
Data table
| Tld | Q3'26 attrition |
|---|---|
| at | 1.9 |
| be | 3.2 |
| ch | 2.5 |
| cz | 3.3 |
| de | 2.6 |
| dk | 2.6 |
| es | 4 |
| fi | 3.7 |
| fr | 4.2 |
| hu | 3 |
| it | 3.1 |
| nl | 3.1 |
| no | 2.6 |
| nu | 4.2 |
| pl | 7.1 |
| pt | 5.4 |
| se | 4.1 |
| sk | 4 |
| uk | 4.3 |
Commercial Insight
What hosting operators should do next
Benchmark your churn performance
Send us your brand and we’ll return your quadrant and top migration corridors for the last 3 months, by TLD.
Terminology & Taxonomy
What does "churn" mean in domains & hosting?
Treating every domain deactivation as "churn" obscures why revenue leaves, and what to do about it. To remedy service cancellations at scale, hosting executives must triage through a combination of internal and external data sources.
| Measure | Evidence source | What it captures | What it does not establish on its own |
|---|---|---|---|
| Customer or account churn | Billing & CRM | A customer account or contractual relationship ends, according to the provider's own definition | Which domains or services moved, where they went, or why the customer left. |
| Received Churn | Product & Control Panel logs | A customer initiates contract cancellation or requests a domain transfer code. May include cancellation survey data. | Only available to the provider for its own services; responses may be biased. |
| Structural Churn: Domain attrition | External (ShareShift) | A domain previously attributed to a provider is no longer observed in its portfolio, under a defined detection rule. | Whether the customer cancelled, moved to a competitor, stopped using the domain or retained other services. |
| Attrition Rate | Calculated measure | Deactivated domains as a share of Total domains; by TLD or provider. | Isolated to a period; annualized data may be misleading. |
| Passive Churn | Billing & CRM | A customer's account is disabled or deleted by the system post non-payment. | Why or when the service stopped being used. |
| Structural Churn: Service deactivation | Product log or External (ShareShift) | A provisioned service, such as mail or web hosting, stops due to user action or a system intervention subject to non-payment. | Whether the domain is commercially active or has a paying customer behind it. |
| Provider churn: Migration event | External (ShareShift) | A domain's provider of DNS, MX or Web hosting changes between observations | Whether the registration provider or registrant also transferred, nor which services moved with it. |
| Net migration | Calculated measure | Incoming observed migrations minus outgoing observed migrations, relative to a stated portfolio denominator | Total portfolio growth or decline, because new domains and deactivations also affect the portfolio |
| Net portfolio change | Calculated measure | The change between opening and closing portfolio under defined inclusion rules | The separate contributions of acquisition, migration, deactivation, reactivation and attribution changes unless these are reconciled |
| Gross Movement Rate | Calculated measure | Gross migrations as a share of total domains; by TLD or provider. Indicates volatility in a cohort. | Net effect on the portfolio size. |
| Net Movement / Switch Rate | Calculated measure | Net migrations as a share of total domains; by provider only. Indicates effect of growth from taking market share. | Isolated to a period; annualized data may be misleading. |
These events happen at different moments in time, and in different combinations. A customer may change the DNS configuration for a domain in April while a service agreement remains active until the next renewal date. Billing may therefore show no churn in April even though an important technical change has already occurred. In another case, the domain may move while the customer continues paying the original provider for registration or another service.
As this article is based exclusively on external data and calculated measures, it relies on domain attrition for observed domain deactivations or disappearances under ShareShift's measurement rules, and on provider migration for changes in DNS provider attribution.
Neither term should automatically be read as confirmed customer churn.
Measurement Integrity
Methodology and limitations
This report draws on ShareShift's core dataset and scoring pipeline: weekly DNS, HTTP and TLS scans across millions of European ccTLD domain names. The latest telemetry scans were delivered on October 19, 2026. Unless otherwise stated, all data is for the three months prior starting on July 27, 2026.
Full detail on data sourcing, scope, attribution rules, and detectability limits lives at shareshift.io/methodology.
With all isolated datapoints, and churn-related information in particular, technical events must not be presented as a confirmed customer outcome. These events are nevertheless relevant as they reveal real changes that are absent from a provider's own billing records. The precise mapping of interpretations for ShareShift's telemetry follows.
| Observation | Reasonable interpretation | What needs additional evidence |
|---|---|---|
| A domain disappears from a provider's observed portfolio | Domain attrition under the stated observation rules | Whether a single scan failed; the customer cancelled; the domain expired, or the site closed |
| DNS provider attribution changes | An observed provider migration | Whether attached services transferred too |
| A or MX records change around migration | The domain's web or mail configuration changed | Whether either the old or the new service were actively used or paid for |
| A competitor receives incoming migrations | The competitor is gaining domains through observed movements | Whether it gained equivalent revenue and/or customer accounts |