Why Did Our Inbound Leads Drop? How to Find Where They Went

Your inbound leads (demo requests, trials and contact forms from people who found you) are down by a third. Nobody changed the website. Rankings look the same. The campaigns are the same ones that worked last year. And yet the weekly lead report keeps sliding, and someone in the leadership meeting has started asking whether content still works, or whether the budget should move to paid.

Before anyone cuts anything, it helps to know that a lot of inbound drops in 2025 and 2026 have had nothing to do with campaigns. Buyers didn’t stop searching. They started getting answers before they reached your site. In Pew Research Center’s July 2025 study of 900 US adults and 68,879 Google searches, people clicked a search result on 8% of visits when an AI summary appeared, against 15% when it didn’t. And in G2’s 2026 survey of more than 1,000 software buyers, 51% said they now start research with an AI chatbot more often than with Google.

This guide gives you a four-step check you can run this week to find where your leads went, a 15-minute way to read Search Console that separates demand problems from click problems, and a 30-day plan to win the pipeline back. It ends with a worked sample audit you can copy.

What are the signs your inbound drop is real, and where did it break?

Inbound breaks in one of four places: tracking, demand, clicks or conversion. Check them in that order, because each one hides the ones after it. A broken form looks exactly like a demand problem on a lead report.

Most teams look at inbound as one number: leads this month versus last month. That number can fall for completely different reasons, and each reason needs a different fix. If a form stopped syncing to your CRM, you need a developer, not a content plan. If people stopped searching for the problem you solve, more blog posts won’t help. If people still search but stop clicking, the fix is to be part of the answer they read. And if they still click but don’t fill in the form, the problem is on your own page.

Run the checks on the same weeks year over year, so seasonality doesn’t mislead you, and split branded searches (people typing your company name) from non-branded ones (people searching for the problem). Then find the first row in the table that matches what you see.

The thresholds are practitioner rules of thumb, not industry standards. Use them to decide where to look first, not as pass-or-fail scores.

Tracking signs: when the leads never left

Tracking breaks are more common than most teams admit, and they’re the cheapest to fix. A new form plugin, a cookie banner update, a CRM field renamed by someone in sales ops, a landing page rebuilt without the hidden UTM fields: any of these can make leads vanish from the report while real buyers keep filling in forms. The tell is a cliff, not a slope. Real demand changes gradually. A tracking break shows up as a sudden drop on a specific day. Submit a test lead through every form on the site, check that it lands in the CRM with the right source, and compare form submissions in your form tool with conversions in GA4 for the last 90 days.

Search signs: demand versus clicks

If tracking is clean, open Search Console. Two numbers matter: impressions, which tell you how often you appeared in results, and clicks, which tell you how often someone came through. When impressions fall, fewer people are searching for what you rank for, or you’ve lost positions. When impressions hold and clicks fall, people still search and still see you, but something on the results page now answers them first. That difference changes everything you do next.

Site and quality signs: when traffic is fine

If impressions and clicks both held, the leak is on your own pages. Compare the conversion rate from your top landing pages to your demo or trial form, year over year. A new form with extra required fields, a slower page, or a weaker offer can cut form fills without touching traffic. And if lead volume held but pipeline fell, the problem is the quality of the leads, not their number.

Is it SEO, ads or the market? How do I tell from Search Console?

Compare impressions and clicks for the same weeks year over year, with branded queries filtered out. Which one moved tells you the cause: impressions held and clicks fell means AI answers took the click; both fell means demand or rankings dropped; both held means look at your site.

This is the single most useful 15 minutes you can spend before the next leadership meeting. It stops the most common mistake: reading a click problem as a demand problem, and cutting the content that still feeds the answers buyers read.

How to run the read

1. Open the Performance report in Search Console and set the date to the last 3 months compared with the same 3 months last year.
2. Add a query filter that excludes your brand name and its common misspellings.
3. Note total impressions, clicks and click-through rate (CTR) for both periods.
4. Switch to the Pages tab and group pages by type: blog and guides, product and features, pricing and comparisons.
5. Place each group in the 2×2 above. Most SaaS sites find their blog and guides in the top-left box (impressions held, clicks fell), and their pricing and comparison pages close to flat.

What each box means

Impressions held, clicks fell. Buyers still search and you still appear, but they get the answer on the results page or in a chatbot before they reach you. This is the most common pattern on informational content in 2025 and 2026. More of the same content won’t fix it; being part of the answer will.

Impressions fell, clicks fell. Either fewer people search for the topic or you lost positions. Check average position first. If positions held, demand moved: buyers may be describing the problem in different words, or the category itself is cooling.

Impressions held, clicks held. Search isn’t your problem. Go back to the conversion and tracking checks.

Impressions fell, clicks held. You lost long-tail, low-intent queries that rarely converted. Usually safe to ignore.

Check paid search the same way

If you run Google Ads, pull the same split there: impressions, clicks and CTR for your non-brand campaigns, year over year. When paid CTR falls on the same queries as organic, the results page itself has changed, not your ads.

How much are AI summaries cutting our clicks?

On informational searches, often by half or more, and the biggest losses fall on the top-of-funnel content most SaaS blogs are built on: definitions, how-tos and “what is” guides.

Three independent studies, using different methods, point the same way. Pew Research Center tracked the real browsing of 900 US adults across 68,879 Google searches in March 2025. When a search showed an AI summary, people clicked a traditional result on 8% of visits, against 15% when there was no summary. They clicked a link inside the summary itself on just 1% of visits, and they ended their browsing session on 26% of pages with a summary, against 16% without one.

Seer Interactive looked at it from the publisher’s side: 3,119 informational queries across 42 client organizations and 25.1 million organic impressions. On queries that showed an AI Overview, organic click-through fell from 1.76% in June 2024 to 0.61% in September 2025. Ahrefs, analyzing 300,000 keywords, found that an AI Overview was associated with a 34.5% lower click-through rate for the page ranking first.

These are correlations across many sites, not a forecast for yours, and the exact size of the gap varies by query type. But the direction is consistent: for questions an AI can answer in a paragraph, fewer people need to click.

Why bottom-funnel pages hold up better

Not every query loses the same way. “What is a lead scoring model” can be answered in a summary. “Best lead scoring software for HubSpot” or “[your product] pricing” can’t, because the buyer needs to see the page to decide. When you split your Search Console data by page type, you’ll usually find pricing, comparison, alternatives and integration pages losing far less than educational posts. Those are also the pages closest to a demo request.

Where did the buyers go?

Into AI chatbots and review sites, and they still arrive, later and with a shortlist already made. The demand didn’t disappear; the first part of the journey moved somewhere you can’t see in Search Console.

G2’s 2026 AI Search Insight Report surveyed more than 1,000 B2B software buyers and decision-makers in March 2026. Half of them, 51%, said they now start their research with an AI chatbot more often than with Google, and 71% rely on AI chatbots somewhere in the software research process. Google hasn’t gone away: 80% still use it somewhere in the buying journey. But it has become one stop among several rather than the front door.

The most important number in that report is the last one in the image: 69% of buyers said AI chatbots surfaced information that led them to choose a different vendor than they expected. The shortlist is now being shaped before anyone fills in your form.

That matters because the shortlist decides the deal. In 6sense’s 2025 Buyer Experience Report, a survey of nearly 4,000 B2B buyers, 95% of purchases went to a vendor that was already on the buying group’s Day One shortlist. If AI answers name three vendors for your category and you’re not one of them, the buyers who would have found you through search may never arrive.

Being cited is the new ranking

Seer’s data hints at what that is worth. On queries with an AI Overview in Q3 2025, organic click-through was 0.70% when the brand was cited in the overview, against 0.52% when it wasn’t. Seer is careful to say this is a correlation, not proof that citation causes clicks, but it fits what most teams see: the brands named in the answer keep more of the traffic that’s left.

What gets a brand cited is not mysterious. AI tools pull from pages that answer the question clearly and specifically, from third-party sources that describe the brand in the same terms (review sites, comparison articles, industry publications), and from a consistent category description across all of them.

How do we win back inbound pipeline?

Fix tracking, protect the pages that still get clicks, get cited where buyers now ask, and measure pipeline instead of leads. In that order, because each step makes the next one measurable.

Step 1: Fix tracking and backfill

Test every form, every UTM parameter and every sync into the CRM. If something broke, fix it and backfill the leads you can recover from the form tool’s own records. Until the numbers are right, every other decision is a guess.

Step 2: Shift effort to bottom-funnel pages

Pricing, comparison, alternatives, integration and use-case pages still need a click, because buyers can’t decide from a summary. Make sure each one exists for your main competitors and use cases, loads fast, states pricing or how pricing works, and has a clear next step.

Step 3: Get cited where buyers now ask

Write a plain category description and use it everywhere: your homepage, your product pages, your review profiles and your partner listings. Ask happy customers for detailed reviews on the review sites your buyers use. Answer the questions buyers ask in complete, specific paragraphs that an AI tool can quote. And earn mentions in the comparison articles and industry publications those tools already cite.

Step 4: Ask buyers how they found you

Add a required free-text “How did you hear about us?” field to your demo and trial forms. Software attribution can’t see a ChatGPT conversation, a podcast or a peer recommendation; buyers can tell you. Review the answers monthly.

Step 5: Report pipeline by channel, monthly

Move the inbound report from lead count to pipeline: opportunities and pipeline value by channel and by self-reported source. A falling lead count with steady pipeline is a different story from a falling lead count with falling pipeline.

Worked example: a sample inbound audit

Here’s what this looks like for a fictional Series A company, “Tidewell,” selling HR-compliance software and running its inbound from HubSpot. The numbers are illustrative sample data, built to show the method. They compare the same weeks in two consecutive quarters.

At first glance, Tidewell’s inbound collapsed: demo requests down a third, organic clicks down by more than 40%. The leadership team’s first instinct was to cut the blog budget and move it to paid search.

The diagnosis told a different story. Week 1 found one real tracking break: a pricing-page form had lost its hidden source field in a site update, which accounted for part of the drop. Week 2’s Search Console read showed impressions almost flat (−3%) while clicks fell 43%, almost all of it on educational posts. Pricing and comparison pages had lost very little.

The self-reported field, added in week 3, explained the rest. Nearly one in five demo requests now named an AI chatbot as where they first heard of Tidewell. Those buyers arrived later in their research and with a shortlist, so they converted better, which is why pipeline fell only 11% while clicks fell 43%. Demand held; the clicks moved.

Tidewell kept the blog, refreshed its most-cited guides, built comparison pages for its three main competitors, and started a monthly AI-answer check. The inbound report now leads with pipeline by channel, with lead count underneath.

Should we measure leads or pipeline now?

Pipeline, by channel and by self-reported source. Lead count alone now punishes the channels that work, because fewer clicks can mean fewer but better-informed leads.

When buyers do more research before they contact you, two things happen at once. Lead volume falls, because fewer people need to fill in a form to learn what you do. And the leads that do arrive are further along, because they’ve already compared options in a chatbot or on a review site. A report that only counts leads sees the first effect and misses the second.

Add self-reported attribution

Software attribution tracks clicks, and the new research journey has fewer of them. A required free-text “How did you hear about us?” field catches what software can’t: an AI chatbot recommendation, a podcast, a Slack community, a colleague. Keep it free-text rather than a dropdown, so buyers use their own words, and group the answers monthly.

What leadership should see

Three things, every month:

1. Pipeline created by channel, with self-reported source alongside software attribution.
2. Win rate by source, so you can see which buyers close, not just which ones arrive.
3. The AI-answer check: for your top buyer questions, whether you’re named and which sources are cited.

Lead count still belongs in the report. It just shouldn’t be the headline.

Putting it to work: a 30-day inbound recovery plan

You can run the whole plan in four weeks with one owner, usually the head of marketing or demand generation, plus a few hours from whoever owns your website and CRM. The output isn’t a slide deck. It’s clean tracking, a clear diagnosis, and a demand generation report that leadership can trust.

Week 1: Rule out tracking

Submit a test lead through every form. Check that each one lands in the CRM with the right source and UTM values. Compare form-tool submissions with GA4 conversions and CRM records for the last 90 days, day by day, and look for a cliff. Split leads by channel so you know which one fell.

Week 2: Read Search Console

Run the impressions-versus-clicks read: last 3 months against the same 3 months last year, non-branded only, grouped by page type. Place each page group in the 2×2. Do the same for paid search if you run it.

Week 3: Get found where buyers ask

List your bottom-funnel pages and fill the gaps: pricing, top three competitor comparisons, alternatives, key integrations. Tighten your category description and use it everywhere. Ask five happy customers for detailed reviews. Add the required “How did you hear about us?” field to every demo and trial form. Run your first AI-answer check.

Week 4: Report pipeline, not leads

Rebuild the inbound report around pipeline by channel and self-reported source, with lead count underneath. Present the diagnosis to leadership with the 2×2 and the numbers behind it. Book the monthly AI-answer check and report review.

The checklist

• Every form tested; leads land in the CRM with the right source
• Form, GA4 and CRM numbers compared for 90 days; no unexplained cliff
• Search Console read done: non-branded, year over year, by page type
• Paid search checked the same way
• Pricing, comparison and alternatives pages live and current
• Required “How did you hear about us?” field on demo and trial forms
• Monthly AI-answer check booked
• Inbound report leads with pipeline by channel and self-reported source

If you’d rather have someone run this with you, that’s exactly what the free roadmap call is for.

Sources

• Pew Research Center, Google users are less likely to click on links when an AI summary appears in the results (22 Jul 2025) — 900 US adults, 68,879 searches
• Seer Interactive, AIO Impact on Google CTR: September 2025 Update (4 Nov 2025) — 3,119 queries, 42 organizations
• Ahrefs, AI Overviews Reduce Clicks by 34.5% (17 Apr 2025) — 300,000 keywords
• G2, The Answer Economy: 2026 AI Search Insight Report — 1,000+ B2B software buyers, March 2026
• 6sense, 2025 B2B Buyer Experience Report — ~4,000 buyers, Nov 2025