The starting point
Holded is business management software for small companies and freelancers: invoicing, accounting, inventory, projects, and HR in one tool.
When this data starts, in June 2020, Holded already had plenty of organic traffic. Its non-branded pages brought in a lot of clicks from Google every month. But only a tiny share of those visits became a sign-up, and even fewer became a paid subscription.
The traffic was there. It just wasn’t the right traffic, or it wasn’t landing on pages that led anywhere near the product.
So the goal was never “more traffic”. It was more sign-ups and more paying customers from people who didn’t already know Holded. Everything we measured and reported on came back to that.
What we did
1. Set a baseline we could trust
Before changing anything, we needed numbers that told the truth.
- Branded and non-branded search, split. People searching for “Holded” were going to find Holded anyway. Mixing them in would hide the real result, or flatter it. Every metric on this page is non-branded only.
- Sign-ups tied to landing pages. Each sign-up and subscription was linked back to the organic page that started the visit. From then on, every page was judged on the customers it brought in, not its visits.
- A technical audit. Crawling and indexation, page templates, duplicate and thin pages, internal linking, and page speed. The fixes went to the development team as tickets with clear acceptance criteria, and we checked each one after it shipped.
2. Competitor analysis
We looked at two kinds of competitors.
The obvious ones are other invoicing, accounting, and business management tools selling to the same small companies and freelancers.
The less obvious ones matter more for SEO: the sites that actually rank for the searches Holded’s customers make. Often those aren’t software companies at all. They’re blogs, tax and government portals, template sites, and publishers.
For each competitor we looked at:
- Which topics and searches bring them traffic, and which of those Holded didn’t cover at all. That’s the content gap.
- What the top results look like for each target search. The type of page (guide, template, free tool, product page, comparison), how deep it goes, how recent it is, and what the person searching is really trying to get done.
- Which of their pages are built to convert. Templates and free tools that lead into a product, comparison and “alternative to” pages, and feature pages built around a specific task.
- Where their links come from, and which of those sources were realistic for Holded to earn too.
The output was a list of opportunities. Each one was tied to a search, a page type, and a reason Holded could win it.
3. Keyword research and the content model
Next, we mapped the searches in Holded’s market to pages, and gave each one a stage and a customer-fit score.
- Product searches. People looking for a type of software, like invoicing software or an accounting program for freelancers. Less volume, but the highest intent.
- Task searches. People trying to do something the product handles: make an invoice, file a tax form, track stock. This is where templates, tools, and guides that show the product doing the job work best.
- Learning searches. People trying to understand a topic. Lots of volume, low intent. Only worth it when the topic sits right next to something Holded does.
The content model set out which page type serves each group, how those pages link to each other, and where each one hands the reader over to a sign-up.
4. Prioritize with RICE and build the roadmap
There were always more opportunities than time, so we scored each one with RICE:
- Reach. How many people search for it, and how much traffic the page could realistically win.
- Impact. How close the search is to a sign-up. A product search scores far higher than a learning search with ten times the volume.
- Confidence. How sure we were, based on what already ranks and how strong Holded’s existing pages were.
- Effort. The writing, design, development, and links needed to compete with the current top 10.
The highest scores became a roadmap in three-month blocks. Each block mixed four kinds of work: improving existing pages, new content, technical fixes, and off-page work. At the end of every block we looked at what had actually moved sign-ups, and re-scored everything else.
5. Ship the work
- Existing pages first. Pages already on the first or second page of Google were the fastest wins. Better titles and descriptions, content rewritten to beat the current top 10, and internal links from Holded’s strongest pages.
- New content from briefs. Every new page started from a brief: the target search, the intent behind it, what the top results cover, what ours had to add, and where it should link. Writers worked from the brief, and each draft was reviewed before it went live.
- Conversion on organic pages. A visit that doesn’t sign up doesn’t count. We tracked the visit-to-sign-up rate page by page and changed what wasn’t working: clearer calls to action, sign-up prompts placed where the reader is ready for them, and templates and tools that lead straight into the product.
- Technical health. Regular audits made sure new templates and releases didn’t break crawling or indexation.
- Links. We earned links from relevant, authoritative sites to the pages that needed them to compete.
6. Report on business results, not rankings
The monthly report tracked non-branded sign-ups, subscriptions, and new MRR. Rankings and traffic were useful for diagnosing problems, but they weren’t the goal, and they weren’t what we reported against.
Results
Across the period in the charts, compared with where each metric started:
- Monthly non-branded sign-ups grew 12.9×, and peaked at 15.6× the starting level in May 2024. Over 2024 as a whole there were 2.4× as many as in 2021.
- Monthly paid subscriptions grew 16×, and peaked at 22× the starting level in January 2025. Over 2024 there were 4.4× as many as in 2021.
- New MRR from organic grew 3.8× between January 2023 and January 2025. Across 2024 it was close to double 2023.
Growth wasn’t a straight line. Sign-ups dipped through the second half of 2022 before climbing again.
The clearest sign of what changed is the sign-up rate. Non-branded clicks stayed in the same range the whole time, and never really grew beyond where they started. But the share of those clicks that became a sign-up grew 18×. The growth didn’t come from more traffic. It came from the right traffic landing on pages built to convert it.
What made the difference
- Measuring sign-ups and revenue from day one, not traffic or rankings.
- Going after searches with buying intent, even when they had less volume.
- A roadmap re-scored every quarter based on what actually moved sign-ups.
- Treating conversion on organic pages as part of SEO, not someone else’s job.