SEO Funnel Metrics That Actually Matter (and What to Ignore in 2026)

Most SEO reporting I see is a wall of green arrows that tells you nothing about whether the phone rang. Rankings moved, impressions went up, and revenue stayed flat. If your funnel metrics can’t tell you where a real customer dropped off, you’re measuring the wrong things.

Below is how I actually audit a funnel in 2026, from the first impression down to the signed deal. I’ll show the metrics that hold up, the ones that quietly lie to you, and the queries you can run against your own data to check.

Start with the funnel shape, not the numbers

An SEO funnel has roughly five stages. Each one has a metric that matters and a metric that gets reported instead.

• Impressions and ranking (did anyone see you)
• Clicks and CTR (did they choose you over the other nine results)
• Engagement (did the page answer the question)
• Conversion (did they take the action you wanted)
• Revenue and retention (was that action worth anything)

Most audits stop at stage two. That’s why they feel useless. A keyword can rank #1, pull 4,000 clicks a month, and still produce zero pipeline because the search intent never matched a buying moment.

1. Share of voice in the SERP, not just your rank

Position 1 is not what it used to be. Between AI overviews, featured snippets, People Also Ask blocks, and local packs, the classic ten blue links get squeezed. In 2026 you need to know what percentage of the visible SERP is yours, not just where your one URL sits.

Track it manually for your top 20 money keywords. Open the result in an incognito window, note which elements appear above the first organic link, and count how many of them are yours. If you own the AI overview citation and the local pack, your effective visibility is high even if a competitor technically outranks you.

2. CTR relative to position, not raw CTR

A 2% CTR is great at position 8 and terrible at position 1. Compare each page against the expected CTR for its average position in your own data, not against some industry benchmark you found in a blog post.

Here’s a query I run against the Search Console bulk export in BigQuery to spot pages that rank well but underperform on clicks:

-- Pages ranking top 5 with weak CTR
SELECT
 url,
 SUM(impressions) AS impressions,
 SUM(clicks) AS clicks,
 SAFE_DIVIDE(SUM(clicks), SUM(impressions)) AS ctr,
 SUM(sum_position) / SUM(impressions) AS avg_position
FROM `searchconsole.searchdata_url_impression`
WHERE data_date BETWEEN DATE_SUB(CURRENT_DATE(), INTERVAL 28 DAY)
 AND DATE_SUB(CURRENT_DATE(), INTERVAL 1 DAY)
 AND is_anonymized_query = FALSE
GROUP BY url
HAVING avg_position <= 5
 AND impressions > 500
 AND ctr < 0.03
ORDER BY impressions DESC;

Every row here is a page that earns visibility but loses the click. Usually it’s a weak title tag or a snippet that promises the wrong thing. Fix the snippet before you touch the content.

3. Query-to-page match rate

This is the one most audits skip entirely. Pull the queries each page ranks for and ask whether the page was actually written for that query. If a page about commercial office cleaning is ranking for “how to clean grout,” you have an intent mismatch, and the clicks you get will bounce.

Group your queries into three buckets: informational, commercial, and transactional. Then check that each landing page targets the right bucket. A service page should not be your top entry point for how-to searches. That traffic converts at roughly nothing.

4. Engaged sessions, not bounce rate

Universal Analytics died in 2023, so if anyone still quotes you a bounce rate, they’re reading an old report. In GA4 the metric that carries signal is engagement rate and engaged sessions per user. A session counts as engaged if it lasts 10+ seconds, fires a conversion, or views two or more pages.

What I care about: does the organic session reach the conversion step at a rate comparable to your paid or direct traffic? If organic sits at 0.4% and direct sits at 3%, the page is attracting the wrong people or failing to convince the right ones.

5. Conversion rate by landing page, segmented by source

A site-wide conversion rate hides everything. Pull it by landing page and by channel. You’ll usually find that two or three pages carry the whole funnel and the rest are decoration.

-- GA4: organic conversions by landing page (via BigQuery export)
SELECT
 (SELECT value.string_value
 FROM UNNEST(event_params)
 WHERE key = 'page_location') AS landing_page,
 COUNT(DISTINCT user_pseudo_id) AS users,
 COUNTIF(event_name = 'generate_lead') AS leads,
 SAFE_DIVIDE(
 COUNTIF(event_name = 'generate_lead'),
 COUNT(DISTINCT user_pseudo_id)
 ) AS lead_rate
FROM `analytics_123456789.events_*`
WHERE _TABLE_SUFFIX BETWEEN FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE(), INTERVAL 28 DAY))
 AND FORMAT_DATE('%Y%m%d', DATE_SUB(CURRENT_DATE(), INTERVAL 1 DAY))
 AND (SELECT value.string_value FROM UNNEST(event_params) WHERE key = 'medium') = 'organic'
GROUP BY landing_page
HAVING users > 100
ORDER BY leads DESC;

Swap generate_lead for whatever your real conversion event is. If you never set one up, that’s the first thing to fix. No conversion event means no funnel, just vibes.

6. Assisted conversions and the path length

SEO rarely closes on the first visit. Someone reads a guide, comes back a week later, then fills out a form. If you only credit last-click conversions, you’ll cut the pages that started the relationship.

Look at the average number of organic sessions before conversion, and which pages show up most often in those paths. On a lot of local service sites, the blog post is the first touch and the service page is the last. Both deserve credit. Both deserve budget.

7. Local pack and proximity signals

For anyone serving a geographic area, the map pack is its own funnel. Track your rank in the local pack separately from organic, and watch the “discovery” versus “direct” split in your Business Profile insights. Discovery searches (people who didn’t know your name) are the ones that measure reach. Direct searches measure brand.

If you’re working in a dense market like Nassau or Suffolk County, proximity drags hard. A business in Commack will outrank a competitor for “near me” searches a mile away even when the competitor’s site is stronger. That’s normal. It’s why I tell local clients to build pages for the towns they actually serve rather than chasing one impossible “Long Island” keyword. A proper Long Island SEO plan is built around those town-level pages, not a single homepage.

8. Revenue per session, not just conversions

Two pages can convert at the same rate and be worth wildly different amounts. A contact form for a $200 job and one for a $20,000 job look identical in GA4. Connect your CRM to your analytics, even if it’s a manual export once a month, so you can attach real deal values to real sessions.

Once you have that, everything upstream gets clearer. You’ll know which keywords feed the high-value pipeline and which just feed your inbox with tire-kickers.

What to ignore

• Domain authority scores. Third-party metrics, not ranking factors. Useful for a rough comparison, useless as a target.
• Total keyword counts. Ranking for 5,000 keywords means nothing if 4,900 get zero clicks.
• Raw traffic growth. More visitors to pages that never convert is a cost, not a win.
• Any single-month comparison. Seasonality and algorithm updates will lie to you. Use 28-day rolling windows and year-over-year.

Building the reporting so it runs itself

Manual audits are fine once. They’re a waste of time monthly. The fix is a small data pipeline: Search Console and GA4 exports landing in BigQuery, a scheduled query for each metric above, and a dashboard that refreshes on its own. That’s a few hours of setup and then it just runs.

If your funnel crosses from marketing data into your actual application (a booking system, a client portal, a lead-scoring tool), that’s when it becomes a development job. I’ve built enough of these to know the reporting layer is usually the easy part; the hard part is making sure the conversion events fire correctly in the first place. That’s why I tend to handle the tracking and the custom web development together rather than handing one off and hoping the other side wired it up right.

The short version

Rankings are an input. Revenue is the output. Everything between them is the funnel, and the only metrics worth your attention are the ones that show where real people stop moving forward. Fix the biggest drop-off first, then re-measure. That’s the whole method.

FAQ

How many metrics should I actually track?

Five or six, tied to the funnel stages above. If a metric doesn’t change a decision you’d make, cut it. A dashboard nobody acts on is just decoration.

Do I need BigQuery to do this, or can I use the GA4 interface?

The GA4 interface handles the basics fine. You need BigQuery once you want to join search data with CRM revenue or run custom path analysis. Start simple, add it when you hit the wall.

How often should I audit the funnel?

Full audit quarterly, light check monthly. Watch for sudden CTR drops on pages that didn’t change, since that usually means a SERP feature appeared above you or a competitor rewrote their snippet.