The statistics on this site come from opening working photographers’ websites in a real browser and recording what is there. This page explains how we choose the sites, what we record and what the numbers cannot tell you.
The studies
Each study was run on its own, on the sites that could answer its question. An about-page study needs sites with an about page, so sample sizes differ, and the same measure can differ slightly between two studies run on different sites. Where two figures on this site disagree, trust the larger or more recent one.
| Study | Sites | What was read | Measured |
|---|---|---|---|
| Platform speed test | 205 | Homepage, on desktop and on a phone | November 2026 |
| Wedding page check | 205 (108 wedding) | Homepage and up to eight linked pages | November 2026 |
| Photography website statistics | 50 | Homepage, desktop | September 2026 |
| Prices | 51 | Pricing pages, read by hand | September 2026 |
| About pages | 32 | About page | September 2026 |
| Phone performance | 49 | Homepage, simulated phone | September 2026 |
| Portfolio size | 48 | Portfolio page | September 2026 |
| Trends | 47 | Homepage, desktop | September 2026 |
| Homepages | 47 | First screen of the homepage, desktop | August 2026 |
| Why sites do not rank | 44 to 52 per measure | Homepage, desktop | August and September 2026 |
How we choose the sites
The earlier studies used the working photographers featured across our own roundups. The November 2026 studies went wider, in four steps.
- Collect. We started from 80 public pages that list photographers’ websites: the example galleries published by Format, Pixpa, Zenfolio, Showit, Wix and Pixieset, and roundup articles on design and photography blogs. That gave us 760 domains, to which we added the sites in our own lists.
- Filter. We removed anything that was not a working photographer: builders, blogs, directories, stock libraries, camera brands, tour operators and parked domains. We judged each site by its title and description, never by its body text, because a blog post mentioning weddings does not make someone a wedding photographer.
- Identify the platform. We read the platform from the page’s HTML, such as its generator tag or the hosts its files load from. Sites we could not match are grouped as “Custom or other”.
- Draw the sample. We kept every site whose title or description names weddings or elopements in English, German, French, Spanish, Dutch or Italian, plus the sites on our wedding list. Then we added other sites at random to bring each platform up to 35 where enough were available. That produced 208 sites.
Three of the 208 could not be measured in each test, leaving 205 in the speed test and 205 in the page check. They are not exactly the same 205, because a site can fail one test and pass the other. A site counts as wedding when its title or description names weddings or elopements, so some of them also shoot other subjects. We could not identify any Pixieset sites in the sample.
What we record for speed
| Setting | Desktop | Phone |
|---|---|---|
| Screen | 1440×900 | 390×844, 2× pixel density |
| Connection | Not throttled | 9 Mbps down, 2 Mbps up, 170 ms latency |
| Processor | Not throttled | 4× slowdown |
| Wait after load | 3 seconds | 4.5 seconds |
Each site gets a fresh Chromium browser with the cache turned off. We record the first view and do not scroll. Sites were tested in batches of four at a time.
- Page weight is the bytes transferred for that first view. Images that load only after scrolling are not counted.
- Largest Contentful Paint (LCP) is the value the browser itself reports. Google treats 2.5 seconds or less as good.
- Image formats are read from the responses the browser receives.
- Medians describe the middle site in a group, so one very heavy site cannot distort the figure.
What we record on full pages
For the wedding page check we loaded each homepage, scrolled it so that lazy-loaded text appeared, then opened up to eight linked pages: contact, about, pricing, FAQ, testimonials, blog, info and portfolio.
| Item | What counts |
|---|---|
| Contact route | A link to a contact or booking page, an email address, a phone or WhatsApp link, or a form with two or more fields |
| Portfolio, about, blog | A link on the homepage, on the same domain, labeled that way |
| Testimonials | A linked testimonials page, a heading or label such as “kind words”, or a client quote in the text |
| FAQ | A linked FAQ page, or a heading or label on any page we read |
| A published price | A figure in a currency beside words like “from”, “package” or “hours”. Prices for prints, workshops, presets and gift cards do not count |
| Email capture | A newsletter or mailing-list signup, or a short email form that is not a contact form |
We checked these rules by reading samples of the matches and the misses. Testimonials and prices are minimums: a quote inside an image or a price in a PDF cannot be read, and a few prices written in unusual ways are missed.
What the numbers cannot tell you
- They are not a random sample. The sites come from example galleries and roundups, so they lean toward established photographers, and a builder’s own gallery tends to show its better sites.
- They do not show market share. Platform groups were capped by design, so the sample says nothing about how many photographers use each one.
- They are one pass from one place. Each site was measured once, from a single connection in Europe. A site served from far away will look slower than it does to its local visitors, and the phone test is a simulation.
- Small groups are rough. We only compare groups of 10 or more sites. Smaller groups are labeled.
- They do not measure results. We can see what a website contains and how it loads. We cannot see how many inquiries or bookings it brings in.
- The web changes. Every figure is dated, and a site may have changed since.
Why we do not publish a list of sites
We publish totals rather than a table of named sites and scores. One measurement from one location is not a fair verdict on someone’s business. Where we name a photographer in an article, it is to show a specific approach.
Corrections
The first version of our wedding study read only homepage navigation on 19 sites and undercounted testimonials, FAQs and prices. We re-ran it in November 2026 on 108 sites, reading up to nine pages each, and replaced those figures. If you think one of our numbers is wrong, please tell us.
