Guide
How to compare websites side by side
Comparing two websites is easy to do badly. Screenshots side by side tell you about taste. A useful comparison reads the same sources for every site, on the same day, and only calls something a gap when both sides were actually measured.
Three rules that keep the comparison fair
- Same source for both sides. Compare a sitemap count with a sitemap count, a PageSpeed score with a PageSpeed score run the same way. Mixing sources produces differences that are only artifacts.
- Same pages for both sides. Compare homepage with homepage and pricing page with pricing page. A rival's best landing page against your blog post says nothing.
- Missing is not zero. If a site blocks the fetch or has no sitemap, the value is unknown. Treating it as zero invents a gap in your favor or against you.
What to put in the table
These rows cover what a buyer, a search engine, and a marketer each notice first. The right-hand column is where the evidence comes from.
| Row | What to record | Source |
|---|---|---|
| Content inventory | URLs listed in declared sitemaps, grouped by page type | robots.txt and sitemap XML |
| Publishing pace | Blog or guide URLs per month from lastmod dates | Sitemap lastmod or RSS |
| On-page basics | Title, description, one H1, canonical, Open Graph | Page HTML |
| Structured data | Schema.org types found, and whether they validate | JSON-LD and microdata |
| Speed | Performance, accessibility, and SEO scores for the homepage | PageSpeed Insights (mobile) |
| Offer | Free plan or trial wording, printed prices, billing periods | The pricing page |
| Trust | Reviews, badges, address, phone, privacy and terms | Page HTML |
| Stack | Analytics, tag manager, chat, testing, CMS | Page code and response headers |
Search traffic, ranked keywords, and backlinks can be added as a separate block from a third-party provider. Keep that block labeled as estimates.
Questions to ask of each row
- Is the difference large enough to matter, or within normal run-to-run variation?
- Would a buyer notice it before deciding?
- Can a marketing or product change move it within a few weeks?
- What is the smallest version of the fix you could ship first?
A row that passes all four is a fix. A row that fails the first two is trivia, even if the number looks dramatic.
Comparing more than two sites
With three to six sites, look for the pattern rather than the leader. If every rival has a pricing page with a free plan and you do not, that is a category norm you are missing. If only one rival has something, it may be a bet rather than a standard.
Doing it automatically
Bizzik builds this table for your site and up to five others, cites the source of every cell, and leaves unmeasured cells marked as unknown. You can open the sample report to see the layout, or read the website competitor analysis checklist for what each row means in more depth.