Website, custom dashboard and SEO for a platform of therapists and coaches

Summary

Geef Inzicht is a Dutch platform where people with physical and mental health complaints find a therapist or coach. The client wanted a new website, its own dashboard to replace an external forms and follow-up tool, and better visibility in Google. I built the site and the dashboard, created close to 2,800 local landing pages and fixed the technical SEO. Since 2 September 2026, more than 300 experts work in the new dashboard, and four new articles for employers are live.

Key facts

About the client

Geef Inzicht has been running since 2023 and connects people looking for help with more than 300 screened therapists and coaches. Visitors search by complaint, type of therapy and location, or use the Hulpwijzer, a short questionnaire that suggests matching experts. The platform is not a care provider itself.

The brief

Forms and follow-up ran through GoHighLevel, an external customer contact tool, separate from the website. The client wanted everything in one place: experts manage their own profile, chat and appointments, and the team works from one admin dashboard. The site also needed to rank better for “therapy type plus city”.

Starting point

Before changing anything, I compared the sitemap with the published pages and read Search Console. Three things stood out.

29 of the 103 expertise pages were missing from the sitemap, including /psycholoog/ (psychologist). The cause: the SEO plugin had a limit of 3,000 pages, while the site had 3,526. The oldest pages were silently dropped.

Of the 173 pages with FAQ schema (structured data for search engines), 85 had at least one answer that differed from the visible text on the page.

And for employer searches, the site did not exist. Search Console showed zero impressions over 90 days for queries containing “verzuim”, the Dutch word for sick leave (measured 16 September 2026).

What I did

A custom dashboard instead of GoHighLevel

I built a WordPress plugin with two sides. Experts get a dashboard with appointments, chat, profile texts, invoices and calendar sync. The team gets an admin dashboard of 22 screens built from five shared components, so every screen behaves the same way. On 2 September 2026, 304 of the 305 listed experts received an account and chat on their profile. Contact details in the chat stay hidden until there is a paid appointment. What I chose not to build: a chatbot next to the Hulpwijzer, which already does that job. I improved the Hulpwijzer instead: each match now comes with one line explaining why that expert fits.

Technical SEO and schema

I raised the sitemap limit to 15,000, which added 334 pages. For FAQ schema the rule is now: the visible text leads. I rebuilt the schema of 63 pages from that text and removed it from 17 pages with no visible questions. I deliberately keep ratings out of the schema.

Local pages and content for employers

I publish combination pages (a therapy type or complaint plus a location) step by step. Beforehand I check in Semrush whether people actually search for a combination. Where they do not, I do not build a page.

For employers I wrote four articles on sick leave. Each opens with a short answer that stands on its own, so AI answer engines can quote it. It is a test: whether more follow depends on the first measurement.

How I checked it

• I measure as a logged-out visitor and without cache, because logged-in users get different HTML.
• I test dashboard screens in a real browser, with a real case from that day and with the account that has the fewest permissions. Every button gets pressed once.
• Zero is not an answer. When a measurement returns zero, I first check whether it actually ran. On one day, that turned out to be a measurement error five times.
• The sitemap is monitored daily: a check counts all published pages and compares them with the sitemap Google actually receives. Before trusting it, I tested whether it raises an alarm on a simulated gap of 526 pages and on a server error.
• In articles I recalculate every figure separately. That caught two calculation errors before an article went live.
• I checked the homepage review claim (“4.9 based on 2,150 reviews”) against the site’s own data: 4.93 across 2,152 reviews (16 September 2026). It holds.
• Since 1 August 2026, GA4 only measures visitors who accept cookies, roughly half. For organic traffic I therefore rely on Search Console.

Results

• Search Console, 19 August to 15 September 2026 versus the previous 28 days: 62 percent more clicks. Average weekly position moved from 28.1 (week of 3 August) to 21.2 (week of 7 September). Not all of this is my work: the growth also coincides with a Google update from 18 to 21 August.
• Sitemap, 16 September 2026: from 3,092 to 3,426 URLs.
• Employer articles: no results yet. The baseline is set; the first check is in mid December 2026.

What’s next

The dashboard keeps growing with the team’s requests; a Microsoft calendar connection is planned. In mid December I will check whether the employer articles rank, and that decides whether more follow.

FAQ

Do hundreds of local pages still work in Google?

Only where people actually search that way. Check search volume first, and only build pages for combinations people look for.

Why a custom dashboard instead of an off-the-shelf CRM?

Because forms, chat and appointments then go straight to the team, and rules such as hiding contact details can be set exactly as needed.

How do you test whether a new audience is worth it?

With one page and a baseline: here, zero impressions for “verzuim” over 90 days. I measure again at a fixed moment, and only then does the client decide on more pages.