# What counts as a good click rate when you are small?

> Real click rate benchmarks by channel, why a small advertiser's number is not comparable to a published average, and the sample size before yours means much.

Published 2026-10-09 · 6 min read · click-rate-benchmarks, advertising, measurement
Canonical: https://bidsurvivor.space/blog/click-rate-benchmarks-for-small-budgets

---

Every founder who runs a first campaign eventually searches for click rate benchmarks, finds a table of industry averages, compares their own number to it, and draws a conclusion. Almost all of those conclusions are wrong, for two reasons that have nothing to do with the quality of the ad. The first is that published averages are dominated by advertisers who are nothing like you. The second is that at small volumes your own number is not stable enough to compare to anything. (Disclosure: published by [BidSurvivor](https://bidsurvivor.space), which sells advertising and publishes its own click numbers, small as they are.)

## The published numbers, and who they describe

Roughly where the industry sits, from the benchmark studies people actually cite:

| Channel | Typical click-through rate |
| --- | --- |
| Google Search | ~6% across industries ([WordStream](https://www.wordstream.com/blog/2026-google-ads-benchmarks)) |
| Meta feed | ~0.9–1.5% |
| Display / programmatic | ~0.05–0.1% |
| Newsletter sponsorship | ~0.5–2% of opens |
| Cold email | wildly variable, and not a click rate in the same sense |

Two things about that table matter more than the numbers in it.

**Search is not comparable to the rest.** A 6% search click-through rate and a 0.06% display rate are not the same measurement taken on different channels. Search counts a click against somebody who typed a query; display counts it against somebody who was reading something else. The [Nielsen Norman Group has been documenting banner blindness](https://www.nngroup.com/articles/banner-blindness-original-eyetracking/) since 1997, and the display number is that phenomenon expressed as a decimal.

**The average is an average of spenders.** The advertisers making up a published search benchmark are mostly running mature accounts with conversion history, negative keyword lists and years of creative iteration. Your first campaign is not a worse version of that. It is a different thing, measured before any of that exists.

## Why your own number will not sit still

This is the part that makes benchmark-comparison actively misleading at small scale, and it is arithmetic rather than opinion.

Say your true click rate is 3%. Show your ad 200 times. The most likely single outcome is 6 clicks, but 2 happens often and so does 11. Expressed as a rate, the same unchanged ad will show you 1%, 3% and 5.5% on three consecutive days.

Compare any one of those to a published 6% and you will conclude something. The conclusion will be about which day you looked.

The threshold for a rate to mean anything is roughly **15 events** — the same rule of thumb that governs [how many clicks a conversion rate needs](/blog/how-many-clicks-before-a-conversion-rate-means-anything). At a 3% click rate that is 500 impressions before the click rate itself is worth quoting, and at a display-like 0.06% it is 25,000. The lower the rate, the more expensive certainty becomes, which is exactly backwards from how it feels.

## What to compare against instead

Three comparisons that survive small numbers, in order of usefulness.

**Your own line against your own other line.** Two versions of the same ad, same placement, same window. Both numbers are noisy, but the noise is the same shape on both, and a difference large enough to see through it is a real difference. This is why small budgets should hunt for order-of-magnitude gaps rather than 15% improvements.

**Clicks against the thing after the click.** A high click rate on an ad that sends people to a page they bounce from is a cost, not an achievement. The click rate you want is the one attached to the [conversion rate on the page it lands on](/blog/homepage-vs-landing-page), and optimising the first without the second is the most common way to spend more money for the same result.

**One channel's number against the same channel's published number for the same placement.** If a newsletter tells you a previous sponsor got 1.2% of opens, that is a comparison worth making, because it is like-for-like. A generic industry average is not.

## What we publish, and what it is worth

Our own board sits at the small end and reports accordingly. [BidSurvivor](https://bidsurvivor.space) sells two twelve-hour slots a day, and for every slot it publishes two separate counts on the front page and on the brand's own page: how many people opened the card, and how many clicked through to the advertiser's site. Not an estimate, not a range from a media kit — the counts.

The honest framing: those are small numbers, and small numbers are exactly what this article says not to draw conclusions from. What they are good for is the comparison in the first bullet above — your line against your other line, on cold traffic, at a price of zero for an empty slot and $1.00 to take a held one. Every visitor also starts with $100 of house credit before signing up, so an early test costs nothing out of pocket. The reason to publish counts that flatter nobody is that a platform quoting only its winners is describing a different product than the one you would buy; that argument is in [why ad platforms hide performance data](/blog/why-ad-platforms-hide-performance-data).

## A rate you can act on

The practical version, stripped of benchmarks:

1. **Do not compare to a published average** until you have run enough volume that your own number stops moving between days.
2. **Count events, not percentages.** Fifteen clicks is a signal. "4.2%" from a denominator of 47 is a rounding artefact wearing a decimal point.
3. **Test one change at a time**, against yourself, in the same placement.
4. **Attach every click rate to what happened next.** A click rate with no conversion attached to it is a number that can only go up while your business goes nowhere.

If the arithmetic is what you want, [the ad budget calculator](/tools/ad-calculator) runs it for any budget and says plainly whether the budget is large enough for the answer to be trusted. Most are not, and that is the finding rather than a failure of the tool.