Companies Using Zendesk

Zendesk is the largest single technology in the help desk category we track, and the size of the gap is worth stating plainly: 113,696 domains against 6,161 for Help Scout and 5,430 for Freshdesk. In most software categories the leader holds a plurality. Here it holds 58.04%, and the two nearest genuine competitors together account for less than a tenth of what Zendesk does. That is a market where the interesting question is not who wins, but who is left to sell to.

The answer is in the size table, and it contradicts how the product is usually described. Sixty four percent of the companies we can size run fewer than 50 people, and 85% sit under 200. Zendesk is talked about as enterprise service desk software, and the enterprise accounts are real, 2,712 companies above 1,000 employees sit in this base, but numerically the platform is a small business product that large companies also buy. Both facts matter for anyone selling into this install base. A list of Zendesk domains is mostly a list of small teams with a support problem big enough to pay for, and a short list of very large organisations running a public service operation.

The platform's install base is 85% companies under 200 people. The revenue is not, and neither is the switching risk. Those two groups need completely different conversations.

The second thing this data settles is what a Zendesk detection actually means, because the answer changes depending on which signal fired. Our detection reads five different things: script hosts, a JavaScript global, a response header, three cookies, and a DNS record for support email routing. Of the 113,263 domains we currently detect, 26,565 were observed loading Zendesk's own script hosts. That is the cohort where we can say the product is rendering on the page. The rest match on the quieter signals, which prove the organisation has an account and has routed something through it, but do not prove a support widget is live on the site someone is looking at. We treat the full number as a floor on organisations with Zendesk and the script cohort as the census of live deployments, and every company we name on this page comes from the second group.

That distinction is not academic. It changes prospecting. A company matching only on the mail record may have a Zendesk instance behind a login, serving internal IT, with nothing customer facing at all. A company matching on the script host has a help centre that its customers reach, which is a project with a budget line and an owner. If you are selling migration services, integrations or support analytics, the second list converts and the first one wastes calls.

The detection URLs in the named list make the point visually. Upwork answers on support.upwork.com. Ticketmaster runs help.ticketmaster.com. Kaseya, a software company that sells IT service management itself, runs its own partner help desk on helpdesk.kaseya.com. LabCorp, Honda, Logitech, Barnes and Noble and Clarks all follow the same pattern. The subdomain naming is so consistent that it functions as its own signal: when a company decides to publish a help centre, it puts it at help or support, and it keeps it there for years.

The public sector presence surprised us. The states of Colorado and Maine, San Francisco Unified School District, IDEA Public Schools and Georgia Institute of Technology are all in the top forty by company size. Government and education have a structural reason to run public help centres, since their users cannot be turned away and the questions repeat at enormous volume, and both sectors buy on multi year contracts that rarely churn. For a vendor selling anything adjacent to Zendesk, that is the most stable segment in the base.

On competition, the honest reading is that Zendesk is not losing to a rival. We tested every plausible destination against a control, comparing how often a churned Zendesk domain runs a competitor with how often a still active Zendesk domain runs the same competitor. Intercom, HubSpot Chat, Gorgias, Tidio, Freshdesk, Help Scout and Atlassian Statuspage all came back at or below their own base rate, which means domains that dropped Zendesk are no more likely to be running them than domains that kept it. The only corridor that survived the test is LiveChat, and it runs in both directions: 258 domains arrived from LiveChat and 114 left for it. Against an install base of 113,696 that is a rounding error, and we publish it because it is real rather than because it is large.

Seven of eight candidate competitors failed the base rate test. When a platform this size loses a customer, the customer is usually not going to a competitor. They are going quiet.

One corridor we deliberately excluded is worth describing, because it is the kind of artefact that produces confident and wrong analysis. Medallia looked like the strongest destination in the raw data by a wide margin, 441 domains and a rate forty times its base. Checking the cohort dissolved it: 429 of those 435 domains also run Adobe Experience Platform Launch and 391 also run Qualtrics. These are sites sharing one copied tag container, so the three products appear and disappear together as a group. Nobody migrated from a help desk to a survey platform. A shared container churned.

The tech stack overlap is where the commercial reading gets sharper. Shopify sits on 14.84% of Zendesk domains and HubSpot on 13.98%. The Shopify overlap is roughly 16,900 merchants who have both an order volume problem and a demonstrated willingness to pay for support software, which is the cleanest qualified segment on this page. The HubSpot overlap is more interesting strategically, because HubSpot sells a service product of its own and 3,729 of these domains already run HubSpot Chat alongside Zendesk. That is a bundling threat rather than a competitive one, and it will not show up in switching data until a renewal forces the choice.

The last number worth watching is OpenAI, detected on 6,445 Zendesk domains, 5.65% of the base. Right now that reads as companies adding an AI answering layer beside the help centre rather than in place of it. The direction that overlap moves over the next year is the single most useful leading indicator in this category, and it is one we will keep measuring at TechnologyChecker.io as the coverage deepens.