This guide is long. But it will save you from drowning in dashboards full of numbers that do not help you make decisions. Most “metrics you should track” articles give you a list. They do not tell you why a metric matters, when it matters, or what to do when something looks wrong.
In this guide, you will learn which web analytics metrics actually drive decisions — and which ones just make dashboards look busy. We will cover the core metrics, break them down by website type, show you how to spot vanity metrics, and walk through building a dashboard you will actually use. If you are new to analytics, start with what web analytics is and why it matters first.
- Not all metrics matter equally — the right ones depend on your website type and business goals
- For each metric, know four things: what it measures, when it matters, the red flag signal, and what action to take
- E-commerce, SaaS, blogs, and lead gen sites each need different priority metrics — one-size-fits-all lists are useless
- Vanity metrics (total pageviews, follower counts) feel good but do not drive decisions. Actionable metrics connect directly to revenue
- Always segment your data, check sample sizes, and compare the right time periods before drawing conclusions

Why Most Metrics Lists Are Useless
Search for “website metrics to track” and you will find dozens of articles listing 10 or 20 metrics. They say “track your bounce rate” and “monitor your traffic sources.” But they rarely answer the question that actually matters: what do I do when the number changes?
A metric without an action plan is just a number on a screen. Knowing your bounce rate is 65% is meaningless unless you know whether that is good or bad for your type of website, what might be causing it, and what steps to take.
The second problem is that most guides treat all websites the same. They tell an e-commerce store and a B2B blog to track the same metrics with the same priorities. That does not work.
A metric only matters if it changes a decision you make. If a number goes up or down and you do the same thing either way, stop tracking it. It is noise.
Metrics That Matter — By Website Type
Different websites have different goals, so they need different metrics. Here is how priorities break down.
E-Commerce Sites
The goal is revenue. Every metric should connect to the purchase funnel. For the full tracking setup, see GA4 e-commerce tracking with GTM.
| # | Metric | Why It Matters | Red Flag |
|---|---|---|---|
| 1 | Conversion rate | Directly measures purchase efficiency | Below 1% on desktop traffic |
| 2 | Cart abandonment rate | Shows where buyers drop off | Above 75% |
| 3 | Revenue per session | Combines traffic quality and conversion | Declining while traffic grows |
| 4 | Average order value | Measures upsell effectiveness | Flat or declining 3+ months |
| 5 | Traffic source quality | Shows which channels bring buyers | High-cost channel with zero conversions |
SaaS / Software Products
| # | Metric | Why It Matters | Red Flag |
|---|---|---|---|
| 1 | Trial signup rate | Top-of-funnel conversion | Below 2% from organic traffic |
| 2 | Signup-to-activation rate | Shows if signups become real users | Below 20% |
| 3 | Pricing page visits | Indicates purchase intent | High traffic, near-zero signups |
| 4 | Engagement rate | Content quality for top-of-funnel | Below 40% on blog content |
| 5 | Traffic by source | Which channels bring high-intent visitors | 80%+ from one source |
Blog / Content Sites
| # | Metric | Why It Matters | Red Flag |
|---|---|---|---|
| 1 | Engagement rate | Shows if readers interact with content | Below 50% on long-form articles |
| 2 | Avg. engagement time | Measures actual reading depth | Under 30s on 2000+ word posts |
| 3 | Returning visitors (%) | Audience loyalty and stickiness | Below 15% |
| 4 | Organic traffic growth | SEO effectiveness month-over-month | Declining 3 consecutive months |
| 5 | Scroll depth | How far readers actually read | Less than 25% reaching midpoint |
Lead Generation Sites
For proper form tracking setup, see the guide on tracking form submissions with GTM and GA4.
| # | Metric | Why It Matters | Red Flag |
|---|---|---|---|
| 1 | Form submission rate | Primary conversion action | Below 2% on landing pages |
| 2 | Cost per lead | Paid acquisition efficiency | CPL exceeding customer lifetime value |
| 3 | Landing page engagement | Message-traffic match | Below 35% on paid traffic pages |
| 4 | Source quality | Which channels bring converting leads | High volume, zero form fills |
| 5 | Pages per session | Research behavior = higher intent | Below 1.5 for non-bounce sessions |
Bookmark the table for your website type. Use it as a checklist when reviewing analytics each week. Ignore metrics that do not appear on your list — they are noise for your business model.

The Core Metrics Explained
For each metric: what it measures, when it matters, the red flag, and what to do. Understanding how event-based analytics works helps you see how these are calculated in GA4.
Sessions
What it measures: A group of interactions within a time frame. In GA4, a session starts on site open and ends after 30 minutes of inactivity.
Red flag: A sudden 20%+ drop week-over-week. Also, sessions increasing while conversions stay flat — traffic quality is declining.
What to do: If sessions drop, check traffic sources for the declining channel. If sessions grow but conversions do not, segment by source — you are attracting the wrong audience.
Engagement Rate vs. Bounce Rate
What it measures: In GA4, an engaged session lasts 10+ seconds, has 2+ page views, or includes a conversion. Engagement rate = % of engaged sessions. Bounce rate = 100% minus engagement rate.
Red flag: Engagement rate below 40% on pages designed to drive action. Below 45% on blog content suggests content-intent mismatch.
What to do: Segment by traffic source. If organic engagement is low, the page does not match search intent — check queries in Search Console. Check page load speed — slow pages kill engagement before users see content.
Conversion Rate
What it measures: Percentage of sessions that complete a goal — purchase, signup, form submission.
Red flag: Declining while traffic stays stable. A big gap between desktop and mobile rates (e.g., desktop 4%, mobile 0.5%) signals a mobile UX problem.
What to do: Never look at sitewide conversion rate alone. Segment by device, source, and landing page. The fix is almost always specific to one segment.
Do not compare your conversion rate to “industry averages” without matching the definition. A site tracking newsletter signups will naturally have a higher rate than one tracking purchases. Compare against your own historical baseline.
Traffic Sources
What it measures: Where visitors come from — organic search, paid, social, direct, referral, email.
Red flag: Over 70% from a single source. That is a dependency. Also, “direct” traffic growing suspiciously often means broken tracking, not brand awareness. Using a tag manager helps track sources accurately.
What to do: Diversify. If organic dominates, invest in email. If paid dominates, build organic content. If direct is unreasonably high, audit UTM parameters on campaigns.
Here is a reference summary:
| Metric | Red Flag Signal | First Action |
|---|---|---|
| Sessions | 20%+ drop week-over-week | Check traffic sources for declining channel |
| Engagement rate | Below 40% on action pages | Segment by source; align content with intent |
| Conversion rate | Declining while traffic grows | Segment by device and source; test funnel |
| Avg. engagement time | Under 10 seconds sitewide | Improve page speed and above-fold content |
| Pages per session | High pages + low conversion | Run path exploration; fix navigation |
| Traffic sources | 70%+ from one channel | Diversify acquisition; audit UTM tagging |

Vanity Metrics vs. Actionable Metrics
Vanity metrics look impressive but do not connect to business outcomes. Actionable metrics change how you behave when they move.
| Vanity Metric | Why It Is Vanity | Actionable Alternative |
|---|---|---|
| Total pageviews | Inflated by bots, refreshes, multi-page sessions | Engaged sessions per user |
| Social media followers | Does not equal attention or intent | Social traffic conversion rate |
| Total registered users | Includes dead accounts | Monthly active users |
| Email list size | Large lists with low opens waste money | Email click-to-conversion rate |
| Time on site (raw) | Long sessions can mean confusion | Engagement time on key pages |
| Total impressions | Impressions ≠ attention | Click-through rate (CTR) |
| Blog posts published | Output ≠ results | Organic traffic per post |
Vanity metrics are not always useless. Pageviews matter if you monetize through ads. The problem is when vanity metrics become primary KPIs for businesses where they do not connect to revenue.
A simple test: “If this number doubles tomorrow, what specific action will I take?” If you cannot answer in one sentence, the metric is vanity for your business.
How to Read Metrics Without Getting Fooled
1. The Sample Size Trap
A page with 3 out of 10 visitors converting has a 30% rate. That sounds incredible — until you realize 10 visitors is not meaningful. Do not make decisions unless you have at least 100 conversions (not sessions — conversions).
2. The Segmentation Trap
Your sitewide conversion rate is 2.5%. Break it down: desktop 4.2%, mobile 1.1%, tablet 2.8%. The “sitewide” number hides that mobile is broken. Always segment by device, source, new vs. returning, and landing page.
3. The Context Trap
Traffic dropped 30% this week. Crisis? Not if this week includes a holiday. Before reacting, rule out: (1) tracking is broken, (2) seasonal pattern, (3) external event, (4) actual site change.
4. The Comparison Period Trap
Comparing Monday to Sunday always looks different. Compare same day of week, or use year-over-year to control for seasonality.
When in doubt, use year-over-year comparison as your default. It automatically controls for seasonality, holidays, and day-of-week effects.

Building Your Own Metrics Dashboard
Default dashboards show too much. Build a custom one in five steps:
- Define your goal — the primary action you want visitors to take
- Pick 5–8 metrics — use the website-type tables above
- Add key segments — at minimum: device type and traffic source
- Set baselines — 4–8 weeks of historical averages. Flag deviations above 15%
- Set a review cadence — weekly quick scan (10 min), monthly deep dive (30–60 min), quarterly strategic review
| Category | Metric | Baseline | This Week | Change | Action? |
|---|---|---|---|---|---|
| Volume | Sessions | [your baseline] | — | — | Yes / No |
| Volume | New users | [your baseline] | — | — | Yes / No |
| Quality | Engagement rate | [your baseline] | — | — | Yes / No |
| Conversion | Conversion rate | [your baseline] | — | — | Yes / No |
| Conversion | Goal completions | [your baseline] | — | — | Yes / No |
| Source | Top channel (%) | [your baseline] | — | — | Yes / No |
| Trend | Sessions WoW | — | — | — | Flag if >15% |

Common Mistakes
- Tracking everything, analyzing nothing. Having 50 metrics does not help if nobody reviews them. Track less, analyze more.
- Treating all traffic as equal. 1,000 visitors from a targeted email are worth more than 10,000 from an irrelevant Reddit post. Segment by source.
- Reacting to daily fluctuations. A single bad day is not a trend. Wait for at least a full week of consistent data.
- Ignoring mobile. Mobile is 50–70% of traffic on most sites. If you never check the mobile segment, you miss the majority of user experience.
- Never auditing tracking. Tags break. Events stop firing. If you have not audited in 6 months, some of your data is likely wrong.
- Benchmarking against wrong competitors. Compare against your own historical data first. Industry benchmarks are directional at best.
If your analytics shows a conversion rate that seems too good to be true, check your tracking before celebrating. Duplicate event firing and misconfigured goals are the most common causes of inflated rates.
Wrap-Up
Metrics are only useful if they change the way you act. The goal is not to track more — it is to track the right things and connect every number to a specific decision.
- Identify your website type and pick your top 5 priority metrics
- Build a focused dashboard with 5–8 metrics maximum
- Set baselines using 4–8 weeks of historical data
- Schedule your review cadence
- Audit your tracking setup to make sure data is accurate
Stop chasing numbers. Start chasing decisions.
Ready for next steps? Learn what a conversion is and how to track it, then define KPIs for your specific business type. If you need help with the technical side, see our guide on tag managers.
Frequently Asked Questions
What is the single most important web analytics metric?
It depends on your website type. For e-commerce: conversion rate segmented by device and source. For content sites: engagement rate plus returning visitor percentage. For lead gen: form submission rate. The most important metric is the one most directly connected to how your website generates value.
How often should I check my analytics?
Weekly for a quick scan (10 minutes). Monthly for deeper analysis with segmentation (30–60 minutes). Quarterly for strategic review. Daily checking usually leads to overreacting to normal fluctuations — except during campaign launches, where daily monitoring makes sense for 1–2 weeks.
Is bounce rate still useful in GA4?
Bounce rate exists in GA4 but is redefined as 100% minus engagement rate. An engaged session lasts 10+ seconds, has 2+ pages, or includes a conversion. This makes GA4 bounce rate more meaningful than the old version. Use engagement rate as your primary metric.
How do I know if my conversion rate is “good”?
Compare against your own historical data first. As a rough guide: e-commerce purchase rates typically range 1–4%, SaaS trial signups 2–7%, lead gen form submissions 2–5%. If your rate is significantly below after segmenting by high-intent traffic, investigate your funnel for friction.
What should I do when a metric suddenly changes?
Follow this order: (1) Check if tracking is broken — tags stop firing, filters get misconfigured. (2) Check for external factors — algorithm updates, holidays. (3) Check for site changes — deployments, new designs. (4) Segment the data — is the change affecting all traffic or just one source/device? The answer almost always lives in the segments.
Need Help With Your Analytics Setup?
Whether you are implementing GA4, setting up consent management, or building custom tracking solutions, I can help you get it right the first time.