How to Set Up an Upwork Project Catalog That Actually Gets Purchases

A Project Catalog listing is a productized service: a fixed scope, a fixed price, a fixed delivery window, published on your profile so clients can order it directly without posting a job or reading a proposal from you. It sits alongside your regular proposal pipeline as a second acquisition channel - one that runs while you sleep, because a client browsing Upwork at 11pm can buy a catalog listing without waiting for you to respond.

It does not work for every kind of freelancer. If your work varies wildly by client - custom software architecture, bespoke brand strategy, anything that needs a discovery call before you can even estimate scope - forcing it into a fixed-price catalog listing produces a bad brief and a client who ordered something different from what you deliver. Catalog works best for writers, designers, video editors, and developers with a deliverable they have built enough times to price without thinking: a 10-page website in Framer, a 90-second product explainer edit, a logo package, a WordPress speed audit.

Look back at your last 10 to 20 completed contracts. Find the one type of job that showed up more than once, took roughly the same amount of time each time, and did not require a long discovery process to scope. That is your catalog candidate. If nothing repeats yet because you are new, pick the single most common request in your niche - the job posting you see over and over when you search your category.

Resist the urge to catalog your full range of services in one listing. 'Website design services' is not a catalog listing, it is a vague promise that forces a client to message you with questions before they can buy anything - which defeats the entire point of Catalog, which is a direct purchase with zero back-and-forth. 'A 5-page WordPress site for a local service business, built and live within 7 days' is a catalog listing. The narrower it is, the less it needs a conversation before purchase.

A catalog listing has the same job as a landing page: convince one specific buyer with one specific problem that this exact offer solves it, in under 30 seconds of reading. Structure it in this order: the outcome the client gets (not the process you follow), exactly what is included, exactly what is not included, the delivery timeline, and the number of revisions. Clients buying without talking to you first need the boundaries spelled out or they will assume unlimited scope and message you asking for changes you never agreed to deliver.

Project Catalog supports up to three pricing tiers per listing. Use all three, and design them so the middle tier is what most buyers actually pick - not the cheapest, not the most expensive. The bottom tier exists to make the middle tier look reasonably priced by comparison; the top tier exists to make the middle tier look like the smart, not-cheap-not-excessive choice.

This is where an AI tool genuinely earns its place in the workflow, not as a gimmick. Open Claude Code as a desktop app (no terminal required), point it at a plain-text file with your scope notes, and ask it to draft three versions of the listing title and first line - one outcome-led, one price-led, one speed-led. Read all three out loud. The one that sounds like something a real buyer would stop scrolling for is your draft. This takes minutes instead of the twenty or so it usually takes to write good short copy from a blank page.

The second use is more useful than the copy draft: ask Claude Code to write you a five-line script that takes your desired take-home per tier and outputs the price you need to list, after Upwork's service fee. You will reuse this script every time you add a new catalog listing or reprice an existing one, and it removes the mental math error that quietly underprices a lot of catalog listings.

Do not publish Claude Code's first draft unedited. It does not know your actual delivery process, your actual revision history with difficult clients, or the specific phrase that makes your niche's buyers trust you. Treat every draft as a starting point you rewrite in your own voice, the same way you would treat a first draft from a junior copywriter.

Catalog is not a replacement for proposals, it is an addition. A catalog listing has no algorithmic boost the way a well-targeted proposal to an active job posting does, and it depends entirely on a client finding your listing through search or your profile. It will not rescue a thin profile or a niche with no real demand - if proposals are not converting because of a positioning problem, a catalog listing built on the same positioning will not convert either.

It also will not protect your Job Success Score any differently than a normal contract - a catalog order that goes badly counts against you exactly the same way a proposal-sourced contract would. Scope it as tightly as you would scope any fixed-price agreement, because the lack of a negotiation conversation before the sale means misaligned expectations show up after the client has already paid.