Guide
Web traffic sources: the main types and which ones pay
Updated · 6 min read · Sources checked Nov 18, 2026
Short answer
Web traffic sources are the places visitors come from before they reach your site. Analytics tools group them into channels: organic search, paid ads, social, referral, email, direct and, more recently, AI assistants, plus affiliate links and tagged campaigns. The source that sends the most visitors is rarely the one that sends the most revenue, so rank sources by revenue per visitor as well as by visits.
Knowing your traffic sources answers where people find you. Knowing which ones pay answers where to spend the next hour or the next dollar. This guide covers the main types, how a tool decides which source a visit belongs to, why some traffic is mislabeled, and how to compare sources on revenue.
The main types of web traffic sources
| Source | What it is | How tools detect it |
|---|---|---|
| Organic search | Unpaid results on Google, Bing, DuckDuckGo and other search engines. | A search engine as the referrer. |
| Paid | Ads on Google, Meta, LinkedIn, Reddit, TikTok and other networks. | utm_medium such as cpc or paid, or an ad click ID like gclid or fbclid. |
| Social | Unpaid posts and profiles on X, LinkedIn, Reddit, YouTube, Instagram and similar. | A social network as the referrer or source. |
| Referral | Links on other websites: blogs, directories, news, partners. | Any other website as the referrer. |
| Newsletters, onboarding emails and announcements. | utm_medium=email, or an email tool as the source. | |
| AI assistants | Citations in ChatGPT, Perplexity, Claude, Gemini and Copilot answers. | An assistant's domain as the referrer, or utm_source=chatgpt.com. |
| Affiliate | Partners paid per sale or per signup. | utm_medium=affiliate or a partner link parameter such as via. |
| Campaign | Tagged links with no referrer, such as QR codes, PDFs and podcast notes. | UTM parameters but no referrer. |
| Direct | Typed URLs, bookmarks and links that lost their referrer. | No referrer and no tags. |
How analytics tools decide where a visit came from
A tool has three signals for each visit, checked in order: tags on the landing URL (UTM parameters and ad click IDs), the referrer the browser sends, and nothing at all. The first rule that matches wins. That's why a tagged link in a newsletter counts as Email even if the reader clicked it in Gmail's web app, and why an untagged link in a PDF counts as Direct.
The referrer usually holds only the domain, not the page. Browsers default to the strict-origin-when-cross-origin policy, which sends only the origin to other sites (MDN). You'll see news.ycombinator.com, not the thread. To know which post or campaign sent a visit, tag the link.
Channel names also differ. In GA4, Google's AI Overviews and AI Mode count as Organic Search, and ChatGPT and similar assistants count as AI Assistant (Google's channel definitions). If you compare two tools, compare the underlying referrers, not the channel totals.
What is referral traffic?
Referral traffic is visits that arrive by clicking a link on another website that isn't a search engine, a social network or an AI assistant. A mention in a blog post, a listing in a directory, a link in documentation or a partner's integration page all send referral traffic.
Referral traffic is usually small and often underrated. A single link from a page your buyers already read, such as an integration directory or a comparison post, can send fewer visits than a viral social post but far more customers. Watch for one false referral: when buyers come back from a hosted checkout like checkout.stripe.com, some tools record the payment page as a new referral and take the credit away from the real source. In GA4 you fix this by listing the checkout domain as an unwanted referral (Google's guide).
Paid traffic sources
Paid traffic sources are ad networks you pay per click, per impression or per result: Google Ads, Meta (Facebook and Instagram), LinkedIn, Reddit, TikTok, X, Microsoft Ads and sponsorships. They're the only sources you can turn up or down on demand, which also makes them the easiest to overspend on.
Judge paid sources on revenue against spend, not on clicks. ROAS is revenue divided by ad spend; your break-even ROAS is the point below which a campaign loses money after costs. Work both out with the ROAS calculator and the break-even ROAS calculator.
Which traffic sources are best?
There's no universal answer, but published benchmarks show how much sources differ. First Page Sage's 2026 report on B2B SaaS companies measures how often a visitor becomes a lead, by channel:
| Channel | Visitor to lead | Lead to MQL |
|---|---|---|
| 2.2% | 38% | |
| SEO | 2.1% | 41% |
| 1.3% | 43% | |
| Webinar | 0.9% | 44% |
| PPC | 0.7% | 36% |
Search converts three times better than paid clicks to a lead in that data set. Your own mix will differ, and benchmarks stop at the lead. What you need is your own ranking by revenue. Here is the live demo ranked the way you should rank yours:
Traffic sources ranked by revenue
Orbit Analytics · synthetic data · last 30 days
Then divide revenue by visitors. Revenue per visitor puts a small channel that converts next to a large one that doesn't on the same scale:
Revenue per visitor by source
Orbit Analytics · synthetic data · last 30 days
How to check your website's traffic sources
- In GA4, open Reports → Acquisition → Traffic acquisition for sessions by channel, or User acquisition for the channel that first brought each user.
- Switch the primary dimension to Session source / medium to see individual sites and tags.
- Pick a period of at least 30 days, so a single launch or post doesn't dominate.
- Export the sources with their key events and revenue, if you send purchase values, and sort by revenue instead of sessions.
- Check the Direct and Unassigned rows. If they're large, tag the links you control (newsletters, social bios, partner links) with UTM parameters.
Large Direct traffic is rarely all bookmarks. It also holds app links, copied URLs, chat messages (dark social) and visits whose referrer was stripped. See direct traffic.
See which traffic sources pay in MetricRaven
MetricRaven sorts every visit into one of nine channels (Paid, Email, Affiliate, AI, Social, Organic Search, Referral, Campaign and Direct) using the rules on Campaign & referral links. Connect Stripe, Paddle, Lemon Squeezy or another provider, and each payment is credited to the visit that first brought the customer, even when they pay days later on a checkout domain.
- Traffic → Acquisition → Channel: visitors and revenue per channel. Click a channel to filter every report to it.
- Acquisition → Referrer and Campaign: individual sites and UTM values, sortable by revenue.
- Insights → Details → Path to purchase, under Channels compared: purchase rate, days to buy and visits to buy per channel (buyer insights).
- Revenue → Paid acquisition: Meta Ads and Google Ads spend next to the revenue each campaign brought (Meta Ads, Google Ads).
For a step-by-step version, read find your best marketing channel.
FAQ
What are the main types of web traffic sources?
Organic search, paid ads, social media, referral links from other websites, email, direct visits and AI assistants. Many tools add affiliate traffic and tagged campaigns as separate groups.
What is referral traffic?
Visits from links on other websites that aren't search engines, social networks or AI assistants, such as blog posts, directories, documentation and partner pages. Your analytics tool shows the referring domain.
What are paid traffic sources?
Ad networks you pay for clicks, impressions or results: Google Ads, Meta, LinkedIn, Reddit, TikTok, X, Microsoft Ads and paid sponsorships. Tools detect them by utm_medium values such as cpc or paid, or by ad click IDs like gclid and fbclid.
Which traffic source converts best?
It depends on your product and audience. In First Page Sage's B2B SaaS data, LinkedIn and SEO visitors became leads about three times as often as paid search visitors. Measure your own revenue per visitor by source before you move budget.
Sources
Figures from these pages were checked on Nov 18, 2026. Charts marked “Live demo data” come from MetricRaven's public demo, which runs on synthetic data, not a real customer's.