Two Different Problems Wearing One Label
Search traffic is a distribution problem; sales leads are a fit problem. Solving the first does nothing for the second, which is why so many companies arrive at the same complaint: the traffic went up and the enquiries did not. Nothing has gone wrong with the search marketing. The two outcomes are governed by different variables, and only one of them can be bought.
Traffic depends on visibility for a query. Leads depend on whether the person behind that query wanted what the page offered, at the moment they arrived, badly enough to identify themselves to a company that will then telephone them. That bar is set almost entirely by the relationship between query, page and offer, not by the volume of traffic.
So a traffic target and a lead target should never be treated as one objective in different units. A programme can hit a traffic number by acquiring visitors who were never going to buy, because those visitors are the cheapest available.
The Arithmetic From Impression to Opportunity
Write the chain out with your own numbers before diagnosing anything. Impressions produce clicks; clicks produce sessions on a page; sessions produce enquiries; enquiries produce leads a salesperson accepts; accepted leads produce opportunities; opportunities produce revenue.
Two things become obvious. The chain is multiplicative, so a modest improvement at two adjacent stages beats a heroic improvement at one. And the largest absolute losses are almost never where the attention goes: click-through rate and form conversion rate get the meetings, while the two stages that usually destroy the most value are the match between query and page, and the acceptance rate after the enquiry reaches sales.
Run it backwards too: start from the revenue the company needs, divide by average deal value, win rate and acceptance rate, and you have the number of enquiries the programme owes. In a market with a few thousand qualified buyers, that arithmetic frequently shows the required lead count cannot come from search at all — a finding, not a failure.
Matching the Query to the Offer
A page that ranks for a research query and asks for a demonstration will convert at close to nothing. The visitor asked what something is; the page replied by asking for a meeting with a salesperson. That is not a friction problem. It is a category error about what the visitor came to do.
The rule is easy to state and routinely ignored: the offer must be one step further along than the query, not five. Someone searching for a comparison will accept a comparison document or a specification. Someone searching for a supplier in their city with a part number in the query will accept a quote request, a telephone number or an availability check — and will resent a general brochure.
This is testable in a week. Take your ten busiest organic entry pages, write down the query landing on each, then write down what the page asks for. Where those do not correspond you have found the leak, and it is usually larger than everything else combined.
The Intent Ladder and What Each Rung Deserves
Sorting queries by intent is more useful than sorting them by volume, because intent determines what the page owes the visitor and what it may reasonably ask in return.
| Intent | What the visitor is doing | What the page should ask for |
|---|---|---|
| Informational | Establishing what something is, or whether it applies to them | Nothing. Answer it, and offer the next question |
| Commercial | Comparing approaches, vendors or specifications | A low-commitment exchange: specification, comparison, calculator |
| Transactional | Selecting a supplier, checking availability, requesting a price | Direct contact — quote, call, availability, meeting |
Three implications follow. Informational pages should be judged on assisted outcomes and returning visitors rather than enquiries, and gating them is usually a net loss. Commercial pages are where most business enquiries are actually won, and where most programmes underinvest. Transactional pages deserve every practical convenience — a visible telephone number, a short form, a named person — because friction there is the most expensive friction on the site. The common failure is treating the whole ladder as one audience and putting the same offer everywhere.
The Offer Is the Variable Almost Nobody Tests
Here is where conventional advice goes wrong. Teams test headlines, layouts, button copy and form length exhaustively and almost never test the thing being offered, even though the offer is the largest single determinant of whether a visitor identifies themselves. Changing it requires someone to build something; changing a button requires nobody's permission.
The default business offer is a demonstration or a contact form, which asks the visitor to enter a sales process. On most sites it is the only ask. The alternatives worth testing trade something concrete for contact details without implying a purchase decision: a specification or compatibility document, a technical assessment of the visitor's own situation, or a configuration estimate.
Two disciplines make this a test rather than a guess. Change one offer at a time, on a page with enough traffic to detect a difference. And judge the result on accepted leads rather than form fills, because a weaker ask reliably produces more submissions and sometimes produces worse ones. An offer that doubles form fills and halves acceptance has cost you money.
The page and the form follow from the offer and matter less than it does. One query, one page, one ask: repeat the language of the query in the first line, ask only for the fields that determine the next action, and resolve firmographic detail from a work email address afterwards rather than interrogating the visitor for it. A form demanding company size, budget and timeline before the visitor has read anything is a qualification interview conducted by a stranger. Keep a telephone number and a named contact visible for the people who are ready, because some of your best enquiries never touch a form.
Traffic Problem, Conversion Problem, or Definition Problem
Three different failures produce the same complaint, and the data to separate them is already in your accounts.
A traffic problem: the pages that do convert have healthy conversion rates, but few people reach them. Non-brand impressions for commercial and transactional terms are low, and most visibility is for informational terms or your own brand name. The fix is visibility and paid coverage on terms that matter, not page work.
A conversion problem: you have impressions and clicks on commercial terms, visitors reach the right pages, and they leave. High-intent entry pages show low enquiry rates while the site converts acceptably elsewhere. The fix is query-to-offer matching, then the page, then the form.
A definition problem is misdiagnosed most often: enquiries arrive in reasonable numbers and sales says it is not getting leads. Look at the acceptance rate. If a large share of enquiries are never accepted, or are accepted and never contacted, marketing and sales are counting different things — and adding traffic makes the argument worse. Settle the definitions in writing before touching the media plan.
Measure Quality by Source, Not Volume by Source
Lead volume by source is the most misleading report in business marketing, because the sources producing the most enquiries are reliably the sources asking for the least commitment. Rank sources by what happened to the leads instead: acceptance rate, opportunity rate, deal value, win rate and cycle length.
Two mechanics make that possible. Pass the click identifier into the customer relationship management system with the enquiry and keep it on the record, so the outcome can be joined back to the acquisition. Then send outcomes back to the ad platform: Google documents enhanced conversions for leads as a supplement to the click identifier rather than a replacement, and from 15 June 2026 offline conversion imports and enhanced conversions for leads uploads moved to the Data Manager API and are blocked in the Google Ads API. Any pipeline built on the older endpoint had to be rebuilt.
Expect small samples. With a few dozen opportunities a year, a source's win rate is a story rather than a statistic, so report it with the count attached and treat quarterly quality figures as directional.
When the Honest Answer Is That the Traffic Was Worthless
Sometimes the conclusion is that the traffic should never have been bought, and that no page, offer or form will rescue it.
The signs are consistent. The search terms report shows queries from students, jobseekers, competitors, or countries you do not sell to. The keyword looked commercial but the results page is full of definitions and directories, which tells you what the engine believes the intent to be. Enquiries arrive and sales declines nearly all of them for reasons unrelated to timing. This is a predictable outcome of close-variant matching, which Google documents cannot be turned off for any match type: exact match means same meaning, not identical string.
When that is the finding, stop the spend on those terms, add the negatives, and write in the report that the traffic was not addressable — rather than reporting a cost per lead that implies otherwise. Then ask what the achievable enquiry count is from the terms that remain. A small honest number is a usable planning input; a large dishonest one funds another quarter of the same mistake.
Frequently Asked Questions
Why is my website getting traffic but no leads?
Almost always because the queries bringing the traffic and the action the pages ask for are at different stages: a page ranking for a research question and offering only a demonstration request converts at close to nothing.
Check it directly: list your ten busiest organic entry pages, write down the query landing on each, and write down what the page asks for. Where they do not correspond, that is the cause.
What is a good conversion rate for B2B search traffic?
There is no defensible benchmark, and any figure quoted as an industry standard should be doubted. The rate depends on the intent mix of the queries, what the page asks for, deal value, and whether you count form fills or leads a salesperson accepts. A transactional page and an informational page in the same business will legitimately differ by an order of magnitude, so use your own site as the benchmark.
Should I gate content to get more leads?
Gate commercial material, not informational material. A specification, compatibility matrix, configuration tool or technical assessment can reasonably ask for contact details, because the visitor receives something concrete and the request signals genuine evaluation.
Gating explanatory content usually loses more than it gains: it suppresses the pages that build credibility, and the details it collects often belong to people with no purchasing role.
How do I know whether my paid search leads are any good?
Join the enquiry to its outcome and report by source on acceptance rate, opportunity rate and win rate rather than volume. That means passing the click identifier into the CRM with the enquiry, then sending outcomes back to the ad platform so bidding optimises toward accepted leads rather than form submissions.
Read the search terms report weekly for the first two months.