Loading…

Why this post does not contain a benchmark table

Almost every article on this subject leads with a grid of industries against CTR, CPC and conversion rate. This one does not, and the reason is worth stating plainly rather than hiding.

A benchmark figure is only meaningful if you know the sample it came from: how many accounts, over what period, in which countries, at what spend levels, with which campaign types included. Most published tables disclose none of that. Reproducing numbers whose provenance you cannot describe is how a site loses the credibility it spends years earning, and the numbers are usually wrong for you anyway.

The publishers who do have real data on this, the large agency networks and ad platforms, own these results because they have proprietary panels. Competing with them by restating their figures without attribution is neither honest nor effective.

What an industry average actually hides

Say a benchmark tells you legal has a CPC around six dollars. Inside that average sit personal injury keywords in competitive metropolitan markets at many times that, and low-competition long-tail legal queries at a fraction of it. Your account is somewhere in that distribution, and the average tells you almost nothing about where.

That last row does most of the damage. An account with heavy brand spend will look excellent against any published benchmark while its actual acquisition engine underperforms, because brand traffic was going to convert regardless.

The same problem applies to conversion rate, in a sharper form, which the denominator problem covers.

The comparison that is actually available to you

You have a dataset with none of these problems: your own account history. It is matched on geography, offer, seasonality, and audience by construction, which is more than any published table can claim.

Build the baseline from it. Take the last twelve months, split brand from non-brand, segment by campaign type, and record the distribution rather than a single average. What you want is the range and the median for each segment, because that tells you what normal looks like for you and how much variance is ordinary.

From then on the question stops being whether you beat an industry figure and becomes whether this campaign is above or below your own median for its segment. That question has an actionable answer.

Which of your own numbers are worth keeping in front of you is covered in the metrics worth reporting.

When external benchmarks are genuinely useful

Three cases, and they are narrower than the genre implies.

• Entering a new category. With no history, a rough external figure beats nothing for setting an initial bid ceiling. Treat it as an order of magnitude, not a target.
• Sanity-checking a proposal. If an agency forecasts a cost per acquisition far below anything published for your category, that gap is worth an explanation.
• Budget conversations. A published figure is sometimes the only shared reference point a finance team will accept. Use it, and say out loud what it does not account for.

In none of those cases is the benchmark measuring your performance. It is a prior, and it should be replaced by your own data as soon as you have any.

Organic has a harder version of this problem, because much of its influence never produces a click. Measuring organic when attribution breaks covers it.

What to do when you are genuinely below where you should be

The order matters, because these interact and fixing them out of sequence wastes the learning.

Start with the conversion action itself. A surprising share of accounts diagnosed with a traffic problem are actually measuring conversions incorrectly: duplicated tags, counting a thank-you page reload, or attributing a conversion to the wrong campaign. Confirm the number is real before optimising against it.

Then the landing page, because it sets the ceiling for every campaign pointing at it. Then targeting and match type, which determine who arrives. Only then ad copy, which is the layer most teams start with because it is the easiest to change and the least likely to be the constraint.

To sanity-check what a realistic improvement is worth before committing to it, the SEO ROI calculator models it from your own inputs.