Two minutes from now, you will know which path your site is on.

The two-minute assessment is IIS Rescue's diagnostic for legacy Windows websites. It asks six plain-language questions about one specific site — what the site is built with, what it runs on, what holds its data, whether the source code still exists, what non-web components it depends on, and what the business needs from it next — and recommends one of three paths: rehost, rebuild or replace.

It is free, it asks for no email address or contact details, and the answers stay in the browser. It returns an estimate and the reasoning behind it, never a price: every IIS Rescue project is quoted at one fixed number, in writing, after a free scoping call.

It comes down to where the actual problem lives. If the application still does its job and the thing forcing your hand is the server underneath it, that is a rehost — the site moves as-is onto supported hosting and the code is left alone. If the code itself has become the problem, because it cannot be changed safely or it blocks something the business now needs, that is a rebuild.

The question that settles it fastest is what you need from the site going forward, not what it is built with. “Just get it off the dying server” and “it needs to be a properly modern application again” lead to different paths from identical technology. The two-minute assessment asks that question last and weighs it heavily for exactly that reason.

Each answer carries weight toward one or more of the three paths, and the path with the most weight wins. Older platforms and intact source code lean toward a rebuild being feasible; a site that still works with no appetite for change leans toward a rehost; no database and no application logic leans toward replacing the whole arrangement with something simpler.

A second pass then applies rules that override the raw arithmetic where experience says it would be wrong. Missing source code, for instance, pushes hard against a rebuild rather than toward one. Several unaccounted-for custom components push against replacing the site outright, because nobody can retire what nobody has identified yet.

Nothing about it is generated language. It is a rules engine, so the same six answers always produce the same recommendation, and the reasoning shown on screen is a readout of which rules fired rather than a description of them.

Six: what the website is built with (classic ASP, ASP.NET Web Forms, newer .NET, mostly static pages, or a mix added to over many years); what it is running on; what holds the data; whether the source code still exists; whether there are COM objects, custom IIS modules, ISAPI filters or scheduled tasks behind it; and what you need from the site going forward.

None of them require looking anything up. File extensions usually answer the first one — .asp, .aspx, or plain .html — and if nobody has logged into the server in years, that is itself an answer to the second.

“Not sure” is a valid answer to every one of the six questions, and choosing it is not a failure state. The assessment still returns a path; it just returns it with lower confidence, and it names the specific unknown most worth resolving rather than guessing at it.

Confidence is mostly the proportion of the picture that is actually known, so the questions that move it most are the heavy ones — what the site is built with, whether the source code exists, and what custom components it depends on. Three or more “not sure” answers drops the confidence further still and says so plainly: at that point the result is a starting point rather than a call.

Usually the opposite. With no source code there is nothing to rebuild from except the running site's observable behaviour, which makes a rebuild the expensive path rather than the cheap one. Rehosting what is actually running on the server is generally the sensible move instead.

It does add discovery work, and discovery is the part of a project that moves a fixed price most — which is why it is worth checking whether the code exists before a scoping call rather than during one.

No, but it changes the shape of the job and it is the classic reason a migration stalls. Some COM objects are 32-bit, some are undocumented, and managed hosting platforms will not run them as-is — so they have to be identified before anything can move, and that discovery gets priced in.

It also argues against retiring the software altogether. If components nobody can fully account for are still running, something elsewhere in the business may quietly depend on them, and that needs confirming before the legacy software is switched off.

The two-minute assessment is one, and it will tell you when the answer is no. There is no automated converter that turns classic ASP into modern .NET — VBScript, COM data access and Session state have no mechanical equivalent — so a rewrite is a genuine rebuild whenever it is the right call.

It often is not the right call. Classic ASP still runs on current, supported Windows, so moving the site to supported hosting without touching the code is a legitimate option. It buys time rather than fixing the platform, which is worth deciding deliberately rather than by default.

Because IIS Rescue publishes no rate card, and a tool that guessed at a number from six answers would be guessing at the one thing the practice promises not to guess at. Every project is quoted at one fixed figure, in writing, after a free scoping call, and that figure does not move afterward.

What the assessment does name is what pushes a price up or down on your specific site — missing source code, unidentified components, homegrown data storage — which is the part you can act on before the call.

Accurate enough to plan around, and explicit about when it is not. The result carries a confidence figure derived from how much of the picture the answers actually cover, and it is capped deliberately when the top two paths are close — a genuine toss-up between two paths cannot return a confident answer however few unknowns there were.

At high confidence it says the signals point one way and nothing argues against it. Lower down it names the one thing left to pin down, and at the bottom it says outright that this is a starting point rather than a call. It is an estimate, not a quote: nobody can responsibly commit to a fixed number from six answers on a web page.

Yes — that is what the two-minute assessment is for. It asks for no name, no email address and no company details, nothing is sent anywhere, and there is no follow-up because there is nothing to follow up with. The result is the result.

A free 30-minute scoping call is the next step if you want the fixed price, but it is a step you take when you choose to, with the path already named and the open questions already known.

Yes. The same engine is exposed to AI assistants at iisrescue.com/mcp as a callable tool, so an agent can ask you the six questions and get back the recommendation, the confidence level, the alternatives, the risks and the reasoning, structured for it to relay.

It is the same implementation the page runs, not a second version of it, so an agent and a person cannot be told two different things about the same site. Automated checks compare both faces over a fixed set of scenarios and fail the build on any difference.