4Site's Website Analytics Playbook: Turning Data Into Decisions

---

Build the Measurement Foundation

GA4 configuration and event naming for nonprofits

A default GA4 install tracks pageviews and not much else. That is a starting point, not a measurement program. Before any dashboard or report can be trusted, your event structure needs to reflect what actually matters on your site: form starts and completions, donation amount selections, monthly giving toggles, symbolic giving cart additions, video engagement, resource downloads, whatever your site’s specific conversion paths are.

The most common issue is not missing events, it is inconsistent naming. When “donation_submit,” “Donation Submitted,” and “donate_complete” all exist in the same property because three different people configured tracking over two years, no report built on that data can be trusted without manual cleanup first. Google publishes a set of recommended event names and parameters, and starting from that list rather than inventing your own is the cheapest way to avoid the problem. A documented naming convention, and a single owner who enforces it, is what keeps your event data usable as your team and vendors change.

Consider

• Whether your event naming convention is documented somewhere your whole team and any agency partners can reference before configuring new tracking
• Whether your main conversion actions, including form starts, form completions, donation amount selection, and monthly toggle usage, are tracked as distinct, correctly scoped events
• Whether GA4’s default enhanced measurement events are reviewed and either kept intentionally or turned off, rather than left on by default
• Whether someone audits the event list on a regular cadence to catch drift, duplication, or dead tracking left over from old campaigns

Audit Checklist

• Event naming convention is documented and enforced across internal and agency-configured tracking
• Conversion events are tracked with consistent naming and correct scope
• Enhanced measurement events have been reviewed, not left on default
• A quarterly event audit is scheduled and owned

---

Connecting GA4 to Your Nonprofit CRM

Website analytics that live only in GA4 can tell you about sessions and conversions. They cannot tell you about donor value, retention, or lifetime giving, because that data lives in your CRM. The organizations that get real strategic value from analytics are the ones that connect the two, so a marketing question (“did that campaign work?”) gets answered with a revenue number instead of a traffic number.

This does not require an enterprise data warehouse. At minimum it means UTMs and any relevant tracking IDs are captured on your donation forms and stored against the resulting record in Engaging Networks, Salesforce, or Bonterra, rather than logged in GA4 and then lost the moment the donor converts. If your forms are built on Engaging Networks, our open-source ENgrid framework handles this capture at the template level so it does not have to be rebuilt page by page.

Consider

• Whether UTM parameters or campaign source data are captured as fields on the gift record in your CRM, not just tracked in GA4
• Whether your CRM and GA4 use a shared definition of “conversion” so the two systems’ numbers can be reconciled rather than argued over
• Whether server-side tagging or a Measurement Protocol integration is worth evaluating if client-side tracking is being lost to ad blockers or browser tracking prevention

Audit Checklist

• UTM or source data is captured on the gift record in the CRM, not just in GA4
• GA4 and CRM conversion definitions are documented and reconciled
• Data loss from ad blockers or browser privacy restrictions has been assessed

---

Consent, privacy, and what they mean for your data

Consent is a measurement question before it is a legal one. Once a cookie banner is live, a meaningful share of your traffic stops being measured the moment someone declines, and the size of that gap is not the same on every page or in every state. Teams get caught out when a year-over-year comparison spans the month a banner went up and nobody accounts for the change.

The practical move is to know your number. Compare GA4 sessions against a server-side count or your CRM’s gift count for the same period, write down the gap, and put that figure in the footnote of every report that crosses the date the banner launched. Google’s consent mode lets you keep modeled data flowing for users who decline, which is better than a hole, as long as everyone reading the dashboard knows which numbers are observed and which are modeled.

Consider

• Whether your consent banner’s decline rate is measured and documented, so the reporting gap it creates is a known quantity rather than a surprise
• Whether consent mode is configured, and whether your reports distinguish observed data from modeled data
• Whether session recording tools are configured to mask donor names, email addresses, and payment fields before anyone watches a recording

Audit Checklist

• Consent decline rate is measured and noted in reporting
• Consent mode is configured and modeled data is labeled as modeled
• Personally identifying and payment fields are masked in every behavioral analytics tool

---

Choose metrics that match your goals

Mapping metrics to the five UX lenses

Data without direction is just noise. The fastest way to end up with a dashboard nobody uses is to track everything measurable instead of the specific things that answer your team’s real questions. We recommend anchoring every metric you track to one of the five lenses from the UX Playbook (Findability, Trust, Usability, Persuasion, and Optimize) so that when a number moves, you already know which part of the site experience to investigate.

If organic traffic drops, that is a Findability question. If time on page and return visits fall, that is Trust. If your donation form completion rate drops but traffic holds steady, that is Usability. Building this mapping once means your team spends its time diagnosing problems instead of arguing about which metric matters most.

Consider

• Whether every metric on your primary dashboard is tied to one of the five UX lenses, so a change in the number points to a specific investigation
• Whether your team has agreed on which 8 to 10 metrics actually matter, instead of tracking everything GA4 offers by default
• Whether metric definitions (what counts as a conversion, a session, an engaged session) are documented so reports are comparable over time

Audit Checklist

• Core metrics are mapped to the five UX lenses.
• A defined, limited metric set exists and is documented.
• Metric definitions are written down and consistent across reports.

---

Vanity Metrics vs. Actionable Metrics

Pageviews feel good in a monthly report and rarely change what anyone does on Monday morning. Actionable metrics point directly to a decision: a rising cart abandonment rate on a symbolic giving page tells you exactly where to look. A rising bounce rate on a blog post tells you very little on its own.

The test we recommend, for any metric on your dashboard: if this number moved 20% in either direction, would we know what to do about it? If the answer is no, it belongs in an occasional deep-dive report, not your recurring dashboard.

Sector benchmarks are useful for the same reason. Comparing your donation page conversion rate against the M+R Benchmarks study tells you whether a number that looks bad is actually bad, or just normal for nonprofit traffic.

Consider

• Whether each metric on your recurring dashboard passes the “would we know what to do” test
• Whether metrics are paired with the specific page or flow they describe, rather than reported as site-wide averages that hide where the real problem lives
• Whether your team has retired legacy metrics that were meaningful under an old site structure but no longer map to anything actionable

Audit Checklist

• Recurring dashboard metrics are limited to genuinely actionable ones
• Metrics are segmented by page or flow where relevant, not only reported site-wide
• Legacy or orphaned metrics have been reviewed and retired where appropriate

---

What the Numbers Can’t Tell You: Behavioral Analytics

Session Recording and Heatmaps

GA4 tells you a form’s completion rate dropped. It does not tell you that the monthly giving toggle is overlapping the submit button on mobile Safari. That is the gap behavioral analytics tools fill, with session recordings and heatmaps that show what visitors actually did rather than the aggregate outcome. Microsoft Clarity is free and takes about twenty minutes to install through Google Tag Manager.

The barrier to using these tools well is not cost, it is discipline. Watching five session recordings a week on your highest-traffic donation page will surface more usability problems than a quarter of quantitative reporting alone. Pair that habit with the fundamentals in our guide to building user-friendly donation forms and most of what you find will already have a known fix.

Consider

• Whether a behavioral analytics tool is installed and actively reviewed, rather than installed and forgotten
• Whether session recordings are watched on a regular cadence for your highest-traffic and highest-value pages, particularly donation and symbolic giving flows
• Whether heatmap data is used to test assumptions before a redesign, rather than relying on instinct alone

Audit Checklist

• Behavioral analytics tool is installed and active.
• A recurring cadence exists for reviewing session recordings on priority pages
• Heatmap findings are documented and connected to specific page changes

---

Reading Quick Backs, Rage Clicks, and Scroll Depth

A few specific signals are worth building fluency in, because they surface friction that standard conversion reporting misses entirely. A quick back is a visitor who lands, then immediately returns to search or the referring page, which means the page did not deliver what was promised. Rage clicks, repeated rapid clicks on the same element, usually mean a visitor believes something is clickable and it is not, or a page element is broken. Scroll depth on long-form pages like case studies or annual reports tells you whether your strongest content, the impact numbers and the ask, is actually being seen or living below where most visitors stop.

Consider

• Whether quick back rates are monitored on entrance pages, especially paid and organic landing pages, as an early signal of a mismatch between promise and delivery
• Whether rage click reports are reviewed after any page redesign or new component launch, since they often surface broken interactive elements faster than QA does
• Whether scroll depth data informs where calls to action, impact statements, and the ask are placed on long-form pages

Audit Checklist

• Quick back rate is tracked on priority entrance pages
• Rage click reports are reviewed post-launch for new pages or components
• Scroll depth data has informed at least one placement decision on a long-form page

---

Prove the Value | Attribution and Revenue Tracking

UTM discipline for nonprofit campaigns

UTMs are simple in theory and chaotic in practice the moment more than one person is building campaign links. Inconsistent capitalization, missing parameters, or ad hoc naming (“summer_email” next to “Summer-Email-2026”) fragment your channel reporting into dozens of near-duplicate entries, which makes it impossible to answer “did email or LinkedIn drive more giving this quarter” with any confidence.

A documented UTM naming convention, enforced through a shared template or a link builder like Google’s Campaign URL Builder, is a small process investment that pays off every time someone tries to build a channel performance report. It also makes your email reporting trustworthy, which matters more than most teams realize once deliverability starts affecting who sees the message in the first place.

Consider

• Whether a documented UTM naming convention exists and is used consistently across email, LinkedIn, paid, and partner campaigns
• Whether campaign links are built through a shared tool or template rather than typed freehand by whoever is sending that day’s email
• Whether UTM parameters survive redirects, multi-step donation flows, and any third-party payment processors in the path

Audit Checklist

• UTM naming convention is documented and consistently applied.
• Campaign links are generated through a shared process, not ad hoc.
• UTM data has been verified to survive the full donation flow, including redirects.

---

Revenue Attribution in the CRM

Traffic attribution answers “which channel brought people to the site.” Revenue attribution answers the question leadership actually asks: “which channel brought money to the mission.” Getting there means UTM or source data has to survive the handoff from your website to your CRM and land on the gift record itself, not live in a GA4 report that never talks to Engaging Networks, Salesforce, or Bonterra.

Once that connection exists, you can report revenue by channel and by content instead of sessions by channel. That is the difference between a marketing report and a business case.

Consider

• Whether source and campaign data are captured as fields on the gift record, so revenue-by-channel reporting can run directly from the CRM
• Whether multi-touch attribution, recognizing that a donor may have touched email, LinkedIn, and organic search before converting, is at least discussed even if you start with last touch
• Whether recurring and monthly gift revenue is attributed to its acquisition channel, not just one-time gifts

Audit Checklist

• Gift records include source and campaign attribution data
• Revenue-by-channel reporting is possible directly from CRM data
• Recurring gift attribution is included in channel reporting, not just one-time gifts.

---

Turn Data Into Decisions | Testing and Experimentation

Hypothesis-driven testing on donation pages

Testing without a hypothesis is changing things and hoping. Before any test launches, write down what you expect to happen and why, define the single primary metric that will determine success, and set a guardrail metric. The guardrail is the secondary number you watch to make sure a win on your primary metric is not quietly causing a loss somewhere else, like a shorter donation form that improves completion rate but drops average gift size.

Based on our work with nonprofit clients, a small set of tests consistently moves the needle: donation form structure (single-step against multi-step), call-to-action button language, and auto-populating known donor information for returning supporters. Start there before spending testing effort on lower-impact page elements. Two adjacent ideas worth adding to the backlog early are social proof on the donation page and a one-click second gift on the thank-you page, both of which are cheap to test and easy to measure.

One caution specific to nonprofit traffic: most organizations do not have the volume to reach statistical significance on a small lift in a two-week window. Calculate your required sample size before launch. If the math says the test needs six weeks, run it for six weeks or test something bolder.

Consider

• Whether every test launches with a written hypothesis, a primary metric, and a guardrail metric defined in advance
• Whether tests are prioritized by expected impact and effort rather than by whoever has the strongest opinion in the room
• Whether test duration and sample size are calculated before launch, so results are not called early on data that cannot support the conclusion

Audit Checklist

• Hypothesis, primary metric, and guardrail metric are documented before every test
• A prioritization method exists for the test backlog.
• Sample size and test duration are planned in advance, not decided after the fact.

---

Keeping a test backlog and changelog

A single successful test is a nice result. A living backlog of test ideas, prioritized and owned, is a program. Pair that backlog with a changelog, a running record connecting every material site edit to what you expected and what actually happened, and you build institutional memory that survives staff turnover. It also stops the same test from being quietly re-run two years later because nobody remembers the last result.

Consider

• Whether a prioritized test backlog exists with named owners and target dates, rather than living in someone’s head or a Slack thread
• Whether a changelog connects every material edit and test to its measured outcome, including tests that did not win
• Whether losing and inconclusive tests are documented as thoroughly as wins, since they are often more instructive

Audit Checklist

• Test backlog exists, is prioritized, and has named owners.
• Changelog connects edits and tests to measured outcomes.
• Inconclusive and losing tests are documented, not discarded.

---

Report like a strategist

Dashboard Cadence and Audience

A single dashboard trying to serve everyone usually serves no one well. Your development team needs different information than your executive director, and your executive director needs different information than a board member. Build reporting around cadence and audience: a lightweight weekly or biweekly operational view for the team running tests and fixing pages, and a monthly or quarterly strategic view for leadership that connects site performance to organizational goals like revenue, donor retention, and cost per acquisition.

Consider

• Whether reporting cadence and depth are matched to audience, with operational detail for the team and a strategic summary for leadership
• Whether the leadership-facing report ties web performance explicitly to organizational goals rather than presenting traffic metrics in isolation
• Whether dashboards are reviewed on their stated cadence, not built once and left to go stale

Audit Checklist

• Separate reporting views exist for operational and leadership audiences
• Leadership reporting connects site metrics to organizational goals
• Reporting cadence is defined, documented, and actually followed

---

Framing Data for Non-Technical Stakeholders

The best analytics program in the world does not matter if the people who control budget and priorities cannot act on what it tells them. Leading with the metric and the story behind it, what changed, why it matters, and what you are doing about it, lands better than leading with a chart. Save the underlying chart for backup, not the headline.

Consider

• Whether reports lead with a plain-language finding and recommendation, with supporting charts available but secondary
• Whether every reported problem is paired with a next step or a test already underway, rather than flagged and left open
• Whether non-technical stakeholders are asked what they actually want to know, rather than handed a standard GA4 export

Audit Checklist

• Reports lead with findings and recommendations, not raw charts.
• Every flagged issue has a documented next step and owner
• Stakeholder reporting needs have been directly confirmed, not assumed

---

A 90-day plan to get started

Reading a playbook and running a program are different things. If your analytics setup is currently a default GA4 install and a monthly screenshot, here is the order we would work in.

Days 1 to 30: find out what you actually have. Export your full GA4 event list and mark every event as intentional, duplicate, or dead. Write the naming convention down. Install a behavioral analytics tool and start watching recordings on your main donation page. Measure your consent decline rate so you know the size of your reporting gap.

Days 31 to 60: connect the money. Confirm that UTM and source data reaches the gift record in your CRM and survives every redirect and multi-step form in the path. Reconcile GA4 conversions against CRM gift counts for the same period and document why the two numbers differ. Cut your dashboard down to the 8 to 10 metrics that pass the “would we know what to do” test.

Days 61 to 90: start the program. Build the test backlog and rank it by impact and effort. Run one test on donation form structure with a written hypothesis and a guardrail metric. Set up the two reporting views, operational and leadership, and put the first review on the calendar. Name the owner for measurement architecture, testing, and reporting. It can be the same person, but it cannot be nobody.

---

Frequently asked questions

What should a nonprofit track in GA4 beyond pageviews?

Track the actions that sit on your conversion paths: donation form starts and completions, donation amount selection, monthly giving toggle usage, symbolic giving cart adds, advocacy form submissions, resource downloads, and email signups. Start from Google’s recommended event names so your data stays comparable and future reporting features work without rework.

How do I connect GA4 to Engaging Networks or Salesforce?

At minimum, capture UTM parameters and any tracking IDs as hidden fields on the donation form and write them to the gift record in your CRM. That single step lets you report revenue by channel from the CRM. Server-side tagging and the Measurement Protocol are worth evaluating later, once the basic capture is reliable.

What is the difference between a vanity metric and an actionable metric?

An actionable metric points to a decision. If the number moved 20% in either direction, you would know what to investigate and what to change. A vanity metric, like site-wide pageviews or an aggregate bounce rate, describes activity without telling you where to look or what to do.

How long should we run an A/B test on a donation page?

Long enough to reach the sample size your expected lift requires, calculated before launch, and long enough to cover full weekly traffic cycles. Most nonprofit sites need more than two weeks. Calling a test early on a small sample is the most common way teams end up shipping a change that does nothing.

Do session recordings create donor privacy problems?

They can, which is why masking is a setup step and not an afterthought. Configure your tool to mask names, email addresses, and all payment fields before anyone watches a recording, confirm the masking works on your live donation form, and make sure your privacy policy reflects that recordings are collected.

How often should we report website analytics to leadership?

Monthly or quarterly, in a view built for them. Leadership reporting should connect site performance to revenue, donor retention, and cost per acquisition, and every problem it raises should arrive with a next step attached. Keep the weekly or biweekly operational detail in a separate view for the team doing the work.

---

Where to go from here

Everything in this playbook, the event architecture, the behavioral tracking, the attribution pipeline, the test backlog, the reporting cadence, is part of running a measurement program that informs decisions instead of just existing. That takes ongoing attention: catching event drift before it corrupts a quarter of reporting, keeping UTM discipline consistent as new campaigns launch, and making sure someone is watching the dashboards rather than only building them.
A well-built analytics program drifts without ownership, the same way a well-built website does. If you want a second set of eyes on yours, or help closing the gap between GA4 and your CRM, get in touch. We work with nonprofits on analytics, UX, and fundraising technology, and if we are not the right fit we will give you a fast no and do our best to point you toward someone who is.

---

About the author

Stef Jones is Director of Digital and Web Strategy at 4Site Interactive Studios, where she leads our analytics, UX, and conversion work for nonprofit clients. Her focus is GA4 and Google Tag Manager implementation, donor journey instrumentation, privacy-compliant tracking, and testing programs that connect website performance to fundraising results. Reach her at stef@4sitestudios.com.