Every listing in the Turbo0 directory that carries Similarweb data also carries a traffic-source breakdown — the split between direct visits, search, social, referral links, email, and paid traffic. In theory, that breakdown should let us answer a simple question: do indie tools grow mostly through search, social buzz, referral links, or people just typing the URL back in?
In practice, our data source only answers part of that question honestly, and we'd rather tell you which part than pretend otherwise.
The data problem, up front
Before any chart: in our Similarweb dataset, the search, social, and paidReferrals fields are zero for every single product that has traffic-source data at all — roughly 3,435 items checked, max observed value 0 across all three fields. That is not a finding that indie tools get no search or social traffic. It is a gap in what our current data tier reports. Treating it as "no product is SEO-driven" would be wrong and we're not going to write that sentence.
What does carry real, nonzero values in this dataset is direct, referrals, and mail. So this article sticks to those three, plus a cross-reference to AI-referral traffic from a separate Similarweb AI-platform feed that isn't affected by the same gap. Everything below is scoped to what we can actually verify.
Within Turbo0's indie pool (5,960 alive listings, capped at 5,000,000 monthly visits so giants don't skew the picture), 3,295 products currently have Similarweb visit-history data. Of those, 454 have a nonzero share in at least one of the four dominance-eligible channels (direct, search, social, referrals) — and because search and social are structurally empty, that 454 splits into exactly two groups: 429 direct-dominant and 25 referrals-dominant products.
TL;DR:
- Direct-dominant products (429 of them) have a 30.1% high-growth rate — about 3x the 10.1% baseline across the whole measurable indie pool — and a much higher traffic floor (median 73,559 visits/month vs. 29,821 for referral-dominant products).
- Referral-dominant products (only 25) post an even higher high-growth share (36%), but their median growth is actually negative (-16.1%) — a small, volatile group that swings between spikes and drop-offs rather than a stable trend.
- AI-referral traffic (ChatGPT, Gemini, Claude, Perplexity sending visits) already touches about 5% of the measurable indie pool — a real, distinct third channel worth watching, covered in depth in our AI referral traffic report.
Direct traffic is the strongest signal we can measure
"Direct" means someone typed the URL, used a bookmark, or opened a saved link — no referring page, no search engine, no social platform in between. It's the traffic Similarweb can't attribute to a channel, which sounds like a weakness until you see what correlates with it.
Median monthly visits and high-growth share for the 429 direct-dominant vs. 25 referrals-dominant products in Turbo0's measurable indie pool.
Direct-dominant products aren't just more numerous — they sit on a higher traffic floor (median 73,559 visits vs. 29,821) and grow at nearly 3x the base rate of the general indie pool (30.1% vs. 10.1% high-growth share). We want to be precise about what that is and isn't: it's a correlation inside a snapshot, not proof that direct traffic causes growth. The more plausible read is that direct dominance is a symptom of something else — a memorable name, a product people bookmark and reopen, a workflow tool used daily rather than discovered once. Growth and "people keep coming back on their own" tend to travel together.
Where direct-dominant products cluster
Category tags of the 429 direct-dominant products (products can carry multiple tags, so counts exceed 429).
Platforms (79), Others (73), Image Editing (69), Image Resources (62), and Video Editing (52) lead the list. That roughly tracks Turbo0's overall category mix — bigger categories produce more entries in any cut of the data — but it also fits the pattern above: these are utility categories people return to repeatedly (a background remover, a screen recorder, a note tool) rather than one-time discovery clicks. A case worth naming: Hootsuite, a Management-tagged tool in our dataset, gets 56% of its traffic directly, sits at roughly 2.7M monthly visits, and is growing about 49% month over month — the kind of profile where brand recall is doing real work.
Hootsuite's homepage — a well-established brand name people type back in directly, consistent with its 56% direct-traffic share.
On a much smaller scale, Notesnook, an open-source note app, has one of the highest direct shares in the entire dataset (67%) alongside steady growth — consistent with a tool people save, not search for.
Notesnook's homepage — a privacy-first notes app people bookmark and reopen daily, matching its 67% direct-traffic share, one of the highest in the dataset.
At the extreme end of the growth curve, ExportTok is direct-dominant (30% direct share) and growing about 332% month over month on roughly 874,000 visits — a reminder that "direct-dominant" spans everything from steady brand recall to a tool that's suddenly gone viral through word of mouth alone.
ExportTok's homepage — a single-purpose TikTok comment exporter, growing 332% month over month while still direct-dominant, the "gone viral through word of mouth" end of the spectrum.
The referral paradox: high hit-rate, negative median
Referral-dominant products tell a messier story, and we're presenting it exactly as messy as it is rather than smoothing it over.
Only 25 products in the entire measurable pool are referral-dominant — meaning inbound links from other sites are their single biggest traffic channel. Within that tiny group, 36% qualify as high-growth, the highest share of any segment we measured, higher even than direct-dominant's 30.1%. If you stopped reading there, you'd conclude referral traffic is the best growth channel.
But the median growth rate for this group is -16.1% — meaning the typical referral-dominant product is currently shrinking, not growing. Both numbers are true at once, and the explanation is sample size and volatility, not a contradiction: with only 25 products, a handful of outliers with launch-day or feature-roundup spikes can pull the high-growth share up sharply, while the typical (median) product in the group is watching a one-time referral spike fade back toward its previous baseline as the referring page moves down the feed. This is consistent with what indie makers usually describe about "launch platform" traffic — a Product Hunt feature, a roundup post, or a directory spike gives a real but temporary lift, and once it rolls off, growth can turn negative even while the segment as a whole still contains some genuine breakout stories. The top categories in this group — Others, Platforms, Image Editing, Growth tools, Management — mirror the direct-dominant list closely, which is more evidence this is a small, noisy slice of the same broader population rather than a structurally different kind of product.
The third leg: AI referral, still small but real
Share of the 3,295-product measurable indie pool touched by each channel signal. Categories overlap — a product can be direct-dominant and still receive some AI-referral traffic.
Separately from the dominance split above, 164 products (4.98% of the measurable pool) receive at least some traffic that Similarweb attributes to an AI platform sending visits — chatgpt.com is by far the largest source (averaging 65% of AI-referred traffic across 144 products), followed by gemini.google.com (24% across 85 products), claude.ai (18% across 67 products), and perplexity.ai (9% across 41 products). Five percent is not a wave yet, but it is a real, measurable channel distinct from the zeroed-out search field, and it's growing fast enough that we wrote a dedicated piece on it — see our AI referral traffic report for the full platform breakdown and which categories benefit most.
What this means if you're building an indie tool
Three practical takeaways, scoped to what the data actually supports:
- Optimize for being remembered, not just found. A name people can recall and retype, a favicon they recognize in a crowded bookmark bar, a workflow they return to without re-Googling it — that's what direct dominance looks like from the outside, and it's the one traffic pattern that correlates cleanly with above-baseline growth in this dataset.
- Don't build your roadmap around a single launch spike. The referral-dominant group's negative median growth is a reminder that a Product Hunt feature or a roundup mention is a real but temporary boost. Treat it as a distribution event, not a channel — the goal is converting that one-time spike into people who come back directly next month.
- Email still shows up as a real, if small, signal — and AI referral is the channel to watch next. Mail carries genuine (if modest) traffic in this same dataset, meaning an owned list is still worth building. And with AI platforms now referring measurable traffic to about 1 in 20 measurable indie products, this is a good moment to check whether your own product shows up when someone asks an AI assistant for a recommendation in your category.
Where this data is thin, and why we're saying so
We'd rather under-claim than over-claim. Specifics worth keeping in mind:
- Search, social, and paid-referral traffic are structurally unmeasured in this dataset, not zero in reality. Any product on Turbo0 could be getting meaningful search or social traffic that simply isn't visible to us right now. We're publishing this gap rather than papering over it.
- The referrals-dominant group is 25 products. That's too small for the -16.1% median or the 36% high-growth share to be treated as a stable law of the category — both numbers can move a lot with a handful of additions or removals.
- "Dominant" means largest share among direct/search/social/referrals, not majority share. A product can be classified as direct-dominant with, say, 30% direct traffic if that's simply larger than its other measured shares.
- AI-referral figures come from a separate feed (AI-platform visit attribution) than the direct/referral dominance split, so the ~5% penetration figure and the 429/25 dominance counts are not directly additive — a product can appear in both.
- This is a snapshot, not a longitudinal study. Growth percentages, especially on small visit bases, can look dramatic without representing a durable trend.
If any of these boundaries change how you'd use a specific number, that's the intent — we'd rather you know exactly what's solid and what isn't than take a chart at face value. Browse the underlying Turbo0 category directory to see where your own product would land, and if you're not indexed yet, submit it to Turbo0 so the next traffic snapshot has a data point for you too.
Originally published at turbo0.com.
Top comments (0)