How to tell good traffic from bad
Count leads per visit, source by source. A source that sends 100 visits and 3 calls beats one that sends 1,000 visits and 2. Google Analytics 4 (GA4) does the division for you with a metric called session key event rate:
"The percentage of sessions in which a user triggered a key event." (Analytics Help: Traffic acquisition report)
A session is one visit. A key event is an action you told GA4 counts as a lead: a form sent, a phone tap, a booking. Without key events this column reads zero, so set them up first: GA4 key events for a small business.
Then run four checks, closest to money first:
| Check | Good visitors | Empty traffic | Where to look (free) |
|---|---|---|---|
| Leads per visit | A steady share call, book or send a form | Hundreds of visits, no leads, month after month | GA4: Reports > Acquisition > Traffic acquisition, "Session key event rate" |
| Engagement | Most visits are engaged | Most visits end within 10 seconds on one page | Same report, "Engagement rate" |
| Location | Your service area | Countries or cities you don't serve | GA4: Reports > User attributes > Demographic details |
| What they searched | The service plus a place, a price or "book" | Definitions, job hunting, do-it-yourself, "free" | Search Console: Search performance report; Google Ads: search terms report |
Read them together. One number alone misleads: a visitor who finds your number and calls within 5 seconds looks like a bounce unless the tap is a key event.
At an accounting firm we run marketing for, October's paid ads sent 2,689 visitors and 13 bookings (0.5%). ChatGPT referrals that landed on its service or About pages sent 10 visitors and 2 bookings; Reddit, 8 and 2. From October 1 to November 4, every client who paid came from ChatGPT answers or a Reddit comment, at no ad cost.
What counts as an engaged visit in Google Analytics?
GA4 sorts every visit into engaged or not. A visit is engaged when it "Lasts longer than 10 seconds", "Has a key event" or "Has 2 or more screen or page views" (Analytics Help: Engagement rate and bounce rate).
- Engagement rate is the share of visits that were engaged.
- Bounce rate is the rest: "The bounce rate is the percentage of sessions that were not engaged." Older articles call any one-page visit a bounce; GA4 doesn't.
- Average engagement time per session counts only time your site "was in focus in a user's browser" (Analytics Help: Traffic acquisition report). A tab left open in the background doesn't add to it.
- The 10 seconds is a setting. "The default threshold for engagement time is 10 seconds but can be adjusted" (Analytics Help: Adjust session timeout). If you change it, compare only periods measured the same way.
Google defines these numbers but sets no target for them. Compare your sources against each other and against your own past months.
What does empty traffic look like?
| What you see | Usual cause | What to do |
|---|---|---|
| A burst of visits from one source, seconds long, no leads | Bots and link scanners | Judge sources without that burst |
| Ad visits lasting seconds, from apps and unfamiliar sites | Search ads also showing on the Display Network | Check the campaign's networks and auto-applied recommendations |
| Ad clicks from places you don't serve | Google Ads' default location setting | Switch to "Presence" if you only serve people nearby |
| Searches for definitions, jobs or free how-tos | Broad keywords, or blog posts on side topics | Read your search terms; add negative keywords |
| Forms filled with nonsense | Bots | Add a CAPTCHA to the form |
Bots. GA4 removes the ones it knows:
"In Google Analytics properties, traffic from known bots and spiders is automatically excluded." (Analytics Help: Known bot-traffic exclusion)
You can't see how many it removed, and unknown ones get through. At the same accounting firm, a Google scan after each site release made up 81% of cookieless page views, and LinkedIn's link scanner once counted as 69 fake visitors in a week.
Junk ad clicks. Google calls invalid traffic "any interaction that doesn't come from real people with real interest in an ad." Some of it is accidental, "like when a mobile user reaches for a link and taps an ad instead" (Google Ad Traffic Quality).
Another leak is a Search campaign that also runs on websites and apps. Google's auto-apply recommendations include "Use Display expansion", and once auto-apply is on, "the recommendations will apply regularly" (Google Ads Help: About auto-apply recommendations). At the accounting firm, the Display Network was switched on unnoticed by auto-applied recommendations: in one morning it sent 296 junk-site visits, mostly in Facebook and Instagram in-app browsers, median 5 seconds, against 18 real search visits. GA4's own advice: "if users who come from display are bouncing, make sure your ads are relevant to your site content" (Analytics Help: Engagement rate and bounce rate).
Wrong area. Google Ads' default location option is called Presence or interest:
"This default and recommended option lets you reach people who are likely to be in or regularly present in your targeted location, as well as people who have shown interest in your targeted locations." (Google Ads Help: About advanced location options)
Google suggests Presence when "You only want to target users in specific locations and not users who may be in other locations but are still interested in your product or service." Keep the default if people from elsewhere book you, such as a wedding venue or a mover bringing families into town.
Wrong searches. For free search traffic, Google's free tool that shows which searches bring people to your site and flags pages Google can't index. More shows "breakdowns by queries, pages, and countries" (Google Search Central: Get started with Search Console); see Google Search Console for a small business. For ads, read the The Google Ads report that lists the actual searches that triggered your ads, so you can spot waste. More: Google Ads wasted spend. At the accounting firm, 47–56% of early Google Ads spend went to keywords that matched people looking things up, and only the ad groups aimed at hiring ever produced serious leads.
Fake form fills. Google's advice to advertisers: "Adding a CAPTCHA to your forms will help validate that the person submitting it is human." (Google Ad Traffic Quality: Resources for advertisers)
Is the problem the traffic or the website?
Look at what the engaged visitors do.
- They stay, open several pages, tap "book", then stop. The traffic is fine; the page or offer isn't. See website not getting leads.
- Visitors from one source leave within seconds. That source is the problem. Google: "If a certain channel has a low engagement rate, review your marketing efforts for that channel."
- Every source looks bad at once. Check your tracking first. Google: "If the problem is more widespread, take a look at your measurement code to be sure that all the necessary pages are tagged and that they are tagged correctly." (Analytics Help: Engagement rate and bounce rate) Then check how fast your pages load on a phone: Core Web Vitals.
- Leads come in but few pay. The visits were fine; follow each lead to a paid job: the marketing report to ask for.
Example
Illustrative cases. Numbers are invented for the example.
A landscaper whose visits doubled
Visits rose from 800 to 1,600 a month after a post on lawn diseases started ranking; leads stayed at 12. In Traffic acquisition, the new visits are Organic Search to that one post, mostly from other states, about 15 seconds each, no key events. The 400 monthly visits to service pages bring 11 of the 12 leads. The post costs nothing, but the owner stops reading total visits as success.
A physiotherapy clinic in Calgary on Google Ads
C$1,500 a month buys 600 clicks at a 1% session key event rate: 6 leads. Half the clicks come from mobile apps, because Display expansion was auto-applied. With it off, clicks fall to 300 and the rate rises to 3%: 9 leads for the same budget.
A mover getting forms from overseas
Thirty quote forms a month, a third with made-up names from countries it doesn't serve. A CAPTCHA stops the junk forms. It keeps Presence or interest on its ads, because families moving into its city search from elsewhere.
FAQ
What is a good engagement rate?
Google's help pages define it but set no target. Compare your own sources and months, measured the same way.
Does a high bounce rate hurt my Google ranking?
Google doesn't say so. It uses "aggregated and anonymized interaction data" to judge relevance and names no Google Analytics metric (Google Search: Automatically generating and ranking results).
Do I pay for bot clicks on my Google ads?
Not for invalid clicks Google catches: advertisers "don't have to pay for it" (Google Ad Traffic Quality). Clicks from poor sites aren't fraud, so check where ads ran.
Is more traffic ever bad?
Free visits that don't fit cost nothing, but they drag averages down and hide the sources that work. Paid ones cost you per click.
Different for you?
- Calls, forms and bookings aren't marked as key events yet: GA4 key events.
- Visitors engage but don't call or book: website not getting leads.
- Your ads match the wrong searches: Google Ads wasted spend.
- You want to see which Google searches bring visitors: Google Search Console for a small business.
- Leads arrive but few pay: the marketing report to ask for.
- You'd rather have every lead answered and tracked: Conversion Agent.