The quiet 2.1%.
Shopify's visible storefront population barely moved on net. Underneath that small number, 162,133 storefronts crossed the line in either direction.
The headline is stability. The useful finding is how much movement it took to produce it.
August ended with 14,905 more visible Shopify storefronts than July, a net increase of 2.10%. But 88,519 storefronts entered the CrUX-visible population while 73,614 left it. The market's surface looked calm because two much larger flows nearly cancelled each other out.
This is not a count of businesses opening and closing. CrUX visibility means a storefront received enough real Chrome traffic to enter the public dataset. A store can cross that threshold because its traffic changed, its domain changed, or the business itself changed.
A small net change, made from two large flows.
July and August use the same global HTTP Archive lens: root pages, mobile and desktop, with Shopify detected in the crawl.
91.43% of storefronts that left visibility were in the two longest-tail traffic buckets. Only 768 were in the top-one-million-or-better buckets. Movement is overwhelmingly a long-tail event, not a flight of the largest stores.
Shopify's visible market is not an American market.
Country is read from each shop's own public metadata, not inferred from its domain or currency. The denominator is 696,372 storefronts with a readable country.
Share of storefronts with a readable registered country. Bars are scaled to the United States, the largest row.
Half the market lives between 51 and 1,000 products.
These are true published-product totals from shop metadata, not a first-page estimate. The denominator is 696,394 observed storefronts.
Small catalogs are common, but they are not the whole ecosystem: 39.1% of observed storefronts publish more than 200 products, and 14.6% publish more than 1,000.
Reviews and email own the visible storefront layer.
This is storefront-visible penetration, not total installation share. Checkout-only, admin-only, webhook-only and other apps that leave no public storefront signal cannot appear here.
One storefront can expose several apps, so rows do not sum to 100%. Detection is a lower bound.
The boundaries are part of the result.
- Entered does not mean opened. It means newly visible in CrUX under the same monthly lens.
- Left does not mean closed. It means no longer present above the CrUX visibility threshold.
- App share is a lower bound. Only public storefront signals can be counted.
- No revenue or sales are estimated. This dataset measures observable storefront structure, not business performance.
- Theme and industry rankings are withheld. Their clean coverage is not yet broad enough for a market claim.
The denominator ships with the number.
HTTP Archive monthly crawl, filtered to root pages whose technology list contains Shopify. The CrUX-backed crawl supplies the visibility threshold.
July and August 2026 use the same query lens. August contains 724,513 hosts; July contains 709,608.
Public homepage, /meta.json, /products.json and /collections.json reads. No checkout, account or private merchant data.
Population: August 2026 crawl. Store attributes: latest clean live observation available on 2 September 2026. Counts vary by field and are printed beside each chart.
Population query
SELECT NET.HOST(page) AS host, MIN(rank) AS rank
FROM `httparchive.crawl.pages`
WHERE date = '2026-08-01'
AND is_root_page
AND EXISTS (
SELECT 1 FROM UNNEST(technologies) t
WHERE t.technology = 'Shopify'
)
GROUP BY host;September is already being measured.
Get the next monthly note, or ask for a cut of the data you can put your own store next to.