Choose Plausible when your team needs understandable website measurement with a deliberately limited collection model. Choose Google Analytics when the required workflow includes its broader event and reporting environment. Build a small reporting layer when the real question is which inquiries became qualified opportunities, and that answer lives partly in your intake or CRM system.
The most useful comparison is not “simple versus powerful.” It is whether the measurement you collect supports an actual decision, whether operators understand its limits, and whether the data can be connected to business outcomes without pretending to observe more than it does.
Compare collection, reporting, and business joins
Official sources were reviewed September 26, 2026. The custom column describes Looski's proposed scope. Product descriptions of privacy are not a blanket legal conclusion for every deployment or jurisdiction.
| Criterion | Plausible | Google Analytics 4 | Build your own |
|---|---|---|---|
| Default product focus | Lightweight web analytics with an explicitly privacy-oriented design. Product | Multiple reporting surfaces over event-based analytics data. Reporting surfaces | Collect only events needed for named acquisition and qualification questions. |
| Conversion paths | Funnel analysis is documented. Funnels | Reports and explorations have different capabilities and processing. Surfaces | Define inquiry, qualification, and acceptance as separate outcomes. |
| Revenue | Revenue goals and ecommerce attribution are documented. Revenue | Raw exported events support warehouse joins, with reporting differences. BigQuery | Join outcomes to stable business records; avoid assuming attribution proves causation. |
| Programmatic reporting | Stats API provides programmatic access to analytics measures. API | Data API requests consume quotas based partly on query complexity. Quota guidance | Cache reports, handle incomplete data, and expose freshness. |
| Event submission | Events API can submit pageviews/custom events without the browser snippet. Data access | Exported data reflects events received by Analytics. Export | Use server-confirmed inquiry outcomes alongside browser observations. |
| Warehouse export | Evaluate documented export/API scope for the purchased plan. Access | Standard daily BigQuery export has a one-million-event limit. Export limits | Retain raw business events in your own warehouse where appropriate. |
| Freshness | Test the reporting cadence needed by operators. | Daily and streaming exports have different completeness and cost behavior. Export | Label provisional results and reconcile after late arrivals. |
| Deployment | Hosted product and a self-hosted Community Edition exist; they are different offers. Self-hosting | Managed analytics service with separately operated warehouse resources if used. | Own deployment, collection changes, deletion, monitoring, and backups. |
| Qualified leads | Requires joining analytics context to the business's qualification record. | Requires the same definition and connection to qualification. | Put the qualification decision and reason beside the inquiry. |
| Interpretation | A simpler surface can reduce operational confusion. | Multiple surfaces require careful reconciliation. | Expose the question, definition, missing data, and source record. |
A free collection tool does not make reporting free
For Plausible, use the hosted pricing controls for expected traffic and required features, or separately estimate operating its Community Edition. Self-hosting is not the same as commissioning a new analytics product, and the editions should not be assumed to have identical features. Plausible product and pricing, Community Edition.
For GA4, distinguish the Analytics service from BigQuery storage, processing, and streaming costs. Google documents a standard daily export ceiling and warns that exports can pause when it is exceeded. Streaming has a different behavior and is not a completeness guarantee. These are September 26, 2026 documentation checks, not a workload-specific infrastructure estimate. Export behavior, daily limit.
| Cost or migration item | What to count | Decision consequence |
|---|---|---|
| Collection | Pageviews, custom events, domains, and environments | Remove unnecessary events before projecting recurring usage. |
| Reporting | Operator time and required API/feature access | A simpler report can be worth more than unused breadth. |
| Warehouse | Storage, transformations, queries, and backfills | “Free Analytics” does not cover every downstream resource. |
| Historical migration | Date boundaries, campaign conventions, goals, and filters | Old and new charts may not be directly comparable. |
| Custom business joins | Intake IDs, duplicate handling, and qualification history | This is often the real custom scope. |
| Pilot | Known inquiries from landing through qualification | Reconcile missing and duplicate records before drawing conclusions. |
Buy the view your team will use
Our analysis favors Plausible when a small team needs regular traffic, campaign, and conversion review without a large analytics operating practice. Trial the actual required funnel and revenue features rather than assuming that a minimalist interface lacks them. The current documentation is broader than older “pageviews only” descriptions.
Favor GA4 when its reporting environment and downstream integrations are required and someone owns the configuration. That owner should explain why reports, explorations, API results, and warehouse queries may differ. Google explicitly documents differences between reporting surfaces; a disagreement is not automatically evidence that one query is broken. Reporting comparison.
Neither choice eliminates measurement gaps. Browser restrictions, blocked scripts, missing campaign tags, and offline interactions should be part of the interpretation. Avoid a dashboard that silently turns “unobserved” into “did not happen.”
Build the link from inquiry to useful outcome
For a service business, a focused owned report can connect an inquiry identifier to qualification, appointment, proposal, and accepted work. Keep acquisition observations separate from the staff decision about quality. Store the qualification reason and decision time so a later reclassification is visible.
Deduplicate repeated submissions without deleting their history. A person who sends the same request twice should not automatically count as two new opportunities. Conversely, one person can have two legitimate needs; define the unit of analysis rather than relying on email address alone.
Use campaign context only when it was actually observed and can be carried forward appropriately. Do not fill missing sources with a model's guess. A clear “unknown” category is more useful than precise-looking attribution invented after the fact.
Keep the report narrow: which campaigns produced qualified inquiries, how long qualification took, and where records could not be reconciled. Give operators a way to inspect the underlying records and correct classification with an audit trail. Maintain a distinct view of traffic so a high-converting campaign with very little reach is not mistaken for the largest contributor.
Validate with known inquiries
Trace test cases from arrival to qualification: a genuine inquiry, duplicate submission, rejected request, return visit, and an offline referral entered by staff. Include missing campaign data and a customer who changes the request after submission.
Compare the analytics event, the intake record, and the eventual outcome. Document which links are deterministic and which remain unavailable. Test a failed API request or delayed warehouse update and confirm that the report displays its age instead of presenting partial results as complete.
Measure reporting preparation time and unexplained discrepancies. Use business outcomes to guide decisions while preserving uncertainty about attribution. Buy the collection and exploration you need; build the connection to your own qualification process when that is where the useful answer lives.
Frequently asked questions
Will custom analytics show the exact source of every sale?
No. Attribution has observational limits regardless of who owns the dashboard. Show known evidence and missing data clearly instead of assigning precise credit that the underlying records cannot support.
What would a first custom website analytics include?
Record a small set of first-party events and connect an inquiry reference to later qualification outcomes. Report aggregate counts with explicit definitions and known gaps. Avoid putting personal contact details in analytics events, and document what is collected, retained, and available to each role.
How should we compare the cost of Plausible, Google Analytics, and a custom build?
Include implementation and maintenance even when a product has a free offering. For custom reporting, count collection infrastructure and review of what data is necessary. Compare the same scope and planning horizon, including implementation, retained services, maintenance, and support. A lower subscription bill alone does not establish a lower total cost.