TL;DR: I shipped image → PDF conversion but spent most of the week on SEO content instead of the planned batch UI and landing page. The numbers say that was the right call. Organic search became the #1 traffic source for the first time.
Convertify is a free image converter I'm building solo: Rust + Axum + libvips on the backend, Next.js 16.2 SSG on the frontend, PostgreSQL for landing page content. Every week I publish what worked, what didn't, and the real situation behind it.
Here's week 7.
What actually happened vs. the plan
The plan was clear: ship image → PDF with full batch UI (page size options, orientation picker) plus a dedicated /images-to-pdf landing page, and catch up on community engagement.
What actually happened: image → PDF conversion works — you can convert JPG, PNG, WebP to PDF right now. But the batch UI (combining multiple images into one PDF with layout controls) and the dedicated landing page didn't get built. Instead I went deep on content.
What I did instead of the planned batch work: mass content expansion. Created 9 new landing pages, updated 5 existing ones to full content, improved 5 more that were lagging behind, and found a critical bug in my structured data that had been silently hurting SEO for weeks.
The reason was simple. I opened the database, looked at the gaps, and realized that getting 19 pages from thin/empty to complete would compound for months. The batch UI and /images-to-pdf page would polish one feature. Content would give me 19 new entry points into Google. I picked the compounding bet.
The schema_faq bug — 189 missing questions
This was the real find of the week.
I have faq fields on every landing page (the visible FAQ accordion) and schema_faq fields (the JSON-LD FAQPage markup that Google reads for rich results). They're supposed to match. They didn't.
When I checked: 55 out of 62 pages had fewer questions in schema_faq than in faq. That's 189 questions Google couldn't see in structured data — almost a third of my entire FAQ content was invisible to rich results.
The cause was boring: schema_faq was written manually when I first created each page, and never updated when I added more FAQ entries later. Classic data sync bug.
The fix had two parts. First, a SQL migration that synced all 56 pages. Second, a frontend fix in FAQ.tsx — I removed the landing?.schema_faq ?? fallback so the component now always generates schema from faqItems directly. The desync can't happen again.
This is one of those bugs that costs you nothing visibly but bleeds SEO silently. You don't notice it until you audit.
Content expansion: three clusters
TIFF cluster — 5 new pages. JPG→TIFF, PNG→TIFF, PDF→TIFF, HEIF→TIFF, AVIF→TIFF. These target print/archival workflows — people who need lossless containers for prepress, medical imaging, GIS. Low volume keywords but almost zero competition.
GIF cluster — 4 new pages. WebP→GIF, AVIF→GIF, HEIF→GIF, HEIC→GIF. The use case is mostly "I have a modern format and need an animated GIF for legacy platforms." Slack, older email clients, forums that only accept GIF — it's still a thing in 2026.
HEIF pairs — 5 existing pages rewritten. HEIF→JPG, HEIF→PNG, HEIF→WebP, HEIF→AVIF, HEIF→HEIC. These existed but had minimal content. Now each has 7+ sections explaining the difference between HEIF (the container) and HEIC (HEIF + HEVC codec), with comparison tables and full FAQ.
Partial pages filled — 5 more. PNG→HEIC, HEIC→WebP, GIF→JPG, GIF→BMP, GIF→HEIC all got expanded to full content.
Every page also got cross-links embedded in the content body — teal-styled <a> tags pointing to related conversions. The ContentSections.tsx component now supports dangerouslySetInnerHTML for these, plus paragraph splitting on \n\n so content doesn't render as walls of text.
Frontend work: RelatedConversions upgrade
The internal linking component got a proper overhaul. Before, it was semi-random. Now it uses a VALID_SLUGS set (no more 404 links), reverse-route boosting (if you're on A→B, the B→A page ranks high), family overlap scoring, and a HEIC/HEIF sibling bonus. Max 14 links per page, all validated against actual database slugs.
Small thing, but internal linking is one of those levers where the improvement compounds across every page simultaneously.
The numbers
Google Search Console is moving in the right direction. Indexed pages went from 52 to 59. Impressions had the biggest single-week jump since launch. Average position dropped below 41 for the first time — slow grind but it's grinding the right way. GSC also unfroze after the data lag from week 6, so the dashboard is fresh again.
Clicks are still minimal. The math is straightforward: at position ~40, CTR is basically zero. I need to crack top 20 on at least a few queries before clicks become meaningful. That's not a content problem — that's a time-and-authority problem.
Google Analytics showed the signal I've been waiting for: search.google.com became the #1 traffic source this week — beating Reddit, Threads, and everything else. First time organic search has been on top. Active users grew by 40% week over week, and new users by nearly 70%.
Semrush Site Audit came back clean: 98% health score, zero errors. No broken links, no redirect chains, no orphan pages. The remaining warnings are minor — some pages flagged for low text-to-HTML ratio (expected with Next.js SSG where the JS bundle is heavy relative to static content), and two pages with low word count that I'll bulk up in the next round. The site also passed all Core Web Vitals checks, and the crawl depth stays within 3 clicks for nearly everything. For a solo project with 60+ landing pages, I'm pretty happy with that baseline.
What I'm noticing
1. Content × schema = compound effect. Every page I fix works 24/7 for months. That's fundamentally different from a Reddit post that peaks in 48 hours. The schema_faq fix alone means Google now sees 189 more questions across the site. That's not a one-time bump — it's a permanent improvement to how the site appears in search.
2. Organic traffic is real now. It's still small. But seven weeks ago it was literally zero. The trend line matters more than the absolute number. And when you see Google outpace your referral channels for the first time — that's a signal.
3. Community engagement suffered again. Two weeks in a row with minimal Reddit/dev.to activity. Karma flatlined, HN karma still stuck. This is becoming a pattern I need to break in week 8.
Honest moment
I planned to ship a polished feature with batch UI and a landing page. I shipped the core conversion and then pivoted to content instead. It feels less exciting to write about — "I updated 19 database rows" doesn't have the same energy as "I built a multi-image PDF composer."
But looking at the analytics, this was probably the highest-ROI week so far. Every page I created or fixed is now a permanent surface area for organic discovery. The schema bug fix alone was worth the entire week.
Sometimes the boring work is the important work.
Week 8 plan
Primary: finish the image → PDF experience — batch upload (multiple images into one PDF), page size/orientation options, and the /images-to-pdf landing page. The conversion works, now it needs the polish.
Secondary: re-engage community channels. Daily Reddit/dev.to comments, one new article, one Reddit post. Two weeks of silence is too long.
Not doing: new SEO clusters, big redesigns, or chasing new platforms.
Convertify is a free, ad-free image converter. convertifyapp.net — weekly updates every Sunday.
