A US business funding broker competes for the same searchers as some of the largest financial comparison marketplaces in the country. SEMANTIKOS was asked to do two things: carry out a technical SEO audit of the broker’s website before further marketing investment, and work out which competitor pages were capturing demand in its market.
Lending is a “Your Money or Your Life” category, where Google holds sites to its highest standards for trust, accuracy and expertise. On a site like this, credibility problems are search problems.
What the technical audit found
The audit ran in four passes: code integrity, sitemaps and indexation, real-user Core Web Vitals, and a full crawl. Every finding was re-checked on the live site before it went into the report. The code itself was clean, with no malware, injected spam or black-hat SEO. The problems were structural.
1. Structured data describing content nobody could see
Every page carried the same ten-question FAQ markup, but none of the ten questions appeared on any page. Google’s structured data guidelines require markup to describe content that visitors can see, and this pattern is a common trigger for a manual action. It was almost certainly a removed FAQ widget whose markup kept running. Accurate structured data matters more in finance than in most sectors.
2. Most of the site invisible to crawlers
127 of the 206 URLs in the XML sitemaps had no internal links pointing to them, including core commercial pages. The cause was archive pagination that pointed to page two with a rel="next" tag but no actual link. Google stopped using rel="next" as an indexing signal in 2019, so page two, and everything listed only there, could not be reached by following links.
3. Pages in the index that never should have been
73 indexed pages had no place in search results. More than 50 were single-testimonial pages of around 130 words each. The rest were funnel steps, partner landing pages, consent and document-upload forms, and test pages. Partner pages in search results can also credit organic visitors to the wrong source, which distorts marketing measurement.
4. One root cause behind failing Core Web Vitals
Real-user Largest Contentful Paint was 3.7 seconds against a 2.5-second target. The server alone took 2.1 seconds to respond, leaving almost no time for the page to load. Behind that was a session cookie set on every page for every visitor, which stopped page caching and a CDN from working at all. No front-end optimisation could pass Core Web Vitals while that cookie remained.
5. Content that undermined trust
Two templated sections were 98 to 99% identical, with only a name changed from page to page. One meta description had been copied onto a dozen unrelated articles. A live service page was entirely placeholder text. And two pages quoted different headline figures for the business, the kind of inconsistency that carries more weight in finance than almost anywhere else.
Where competitors were capturing the demand
The second task looked outward. SEMANTIKOS compared the broker’s search visibility with two of the largest US financial comparison marketplaces and a smaller business-loan lead site, using live page checks, around 20,000 competitor keywords and five years of ranking history.
1. A handful of pages capture most of the demand
In each priority topic, one page on each marketplace collected most of the search traffic, and every one funnelled visitors into an application form. They were not always the obvious pages. One marketplace’s biggest lead page in its category was a loan calculator, not an article.
2. Leads are not captured where they rank
One marketplace sent every business-loan application, 19 of the 19 buttons tested, to a separate lending marketplace it owns. The smaller lead site barely ranked in the broker’s market at all, so its leads were coming from outside organic search.
3. Nobody owns the head terms any more
Neither large marketplace ranked in the top 50 for the category’s broad head terms. Both now win on longer, more specific searches, which is exactly where a smaller specialist can compete.
4. A competitor outranking the broker for its own name
A marketplace’s review of the broker ranked first for searches about the broker’s reputation. The page never linked to the broker’s website, and its only calls to action led into the marketplace’s own application forms. People checking the broker’s credibility, often just before applying, were being captured by a competitor.
5. A large and winnable gap
The broker ranked in the top 20 for 74 keywords. The competitors ranked for more than 6,000 business-loan keywords it did not, worth around 1.1 million searches a month. Meanwhile, both large marketplaces had lost around 60% of their top-three rankings since early 2024. Their hold on page one is weaker than their size suggests.
Getting the data right
Keyword tools produce numbers that look precise and can be quietly wrong. Four examples from this project:
- A branded filter that removed real keywords. The keyword tool counted a government agency’s name as a brand. Filtering on it would have silently dropped around 300 relevant keywords.
- Grants inflating a loan category. Grant searches made one category look nearly four times larger than it was. The broker does not offer grants, so they were separated.
- A rank-tracking change in 2025. In September 2025, Google switched off the setting rank trackers used to load 100 results at once, and positions 11 to 20 fell in every tool at the same time. The five-year trend was measured on top-three rankings, which were unaffected.
- Redirects to a bot check. Two links to federal regulations appeared to redirect to a new address, which turned out to be a bot-check page. A bulk find-and-replace would have sent readers to a captcha, the same kind of bot protection found in a 100-million-page directory audit.
The recommended roadmap
- Immediately: remove the hidden FAQ markup, restore links to orphaned pages, and take funnel, form, testimonial and test pages out of the index.
- With the developers: stop the session cookie on anonymous visits, then add page caching, a CDN and a server-level redirect.
- Content: rebuild the templated sections as genuinely distinct pages, create pages for uncovered topics, and publish a trust page that answers reputation searches directly.
- Linking: replace outbound links to comparison sites with primary sources such as government and regulator pages.
- Throughout: route any change to rates, terms or eligibility wording through compliance review.
Why it matters
In lending, technical SEO and trust are the same problem. Markup describing content nobody can see, pages search engines cannot find, and a competitor answering questions about the broker’s own reputation all erode the credibility that rankings in this category depend on. Good financial services SEO is less about chasing keywords than about making a site the most reliable answer for its own market.
Client details have been withheld for confidentiality. This case describes the audit findings and the recommended plan.