How Hard Is It to Sell a Patent? The Real Numbers

How hard is it to sell a patent? Harder than almost anyone tells you, and easier to predict than most people think.

Ask on a forum and you get anecdotes. Ask a law firm and you get "it depends". Ask the people who track the market for a living and you get a number, and the number is uncomfortable: most patents offered for sale never find a buyer.

That is the honest starting point. The useful part comes next, because the same data shows which patents do sell, how long it takes, and which things a seller controls that measurably change the odds.

The short answer: most listed patents never sell

The best public evidence on patent sales comes from the brokered market, the semi-public pool of packages that brokers circulate to buyers. It is only a slice of all patent deals, because many trades happen privately, but it is the slice that is measured.

The measurement is sobering. Brokered packages listed from 2016 to 2020 sold 21% of the time. Earlier analysis of the same market concluded that 67% of packages will never sell.

Read that the right way round. These are packages that a professional broker already agreed to take on, which means someone had already screened them. The raw pool of patents that owners would like to sell does worse.

So the answer is: hard. Roughly four out of five listed packages do not close. But "hard" is not the same as "random", and that distinction is where a seller gets leverage.

What the market data actually shows

Sell-through by listing year

The newest figures, from the 2025 Richardson Oliver patent market report, track what share of each year's listings has sold so far. Packages first listed in 2019 are only now reaching a 30% sold mark. Packages from 2020 sit at 20%. Those from 2021 are at 29%.

Two things follow. A realistic ceiling for a well-marketed package is somewhere around one in three. And reaching that ceiling takes years, not weeks.

The trend has also moved against sellers over the long run. Back in 2014 the same analysts reported that the chance of selling a package had fallen from 51% for 2009 listings to 33% for 2011 listings. The market professionalised, buyers got pickier, and sell-through dropped.

How long a sale takes

The sales that do happen cluster early. In the 2019 analysis, 55% of sales occurred in the first year after listing, followed by a long tail that trickles on for years. For 2020 listings, 60% of the sales happened within eight months of buyers receiving the package, and 30% within four.

Put that together and you get the shape of a real sale: a burst of buyer attention in the first months, then a long quiet period in which a few more packages move, usually at a lower price.

The practical meaning for a seller: your first eight months on the market are the ones that count. If the package is not ready when it goes out, the best window is spent answering questions the buyer should never have needed to ask.

What sold patents fetch against asking

Selling is one question. Selling at your price is another. The Richardson Oliver model estimates closing prices by discounting asking prices by 25% to 55%, depending on how long the package sat on the market. Older packages sell at bigger discounts, and the report tells buyers directly that an old listing is a chance to offer lower.

If you want the numbers on price rather than probability, we cover them in what patents actually sell for.

Why it is harder for an individual inventor

Here is the part that stings for the audience most likely to search this question.

The same 2025 report found that sold packages come disproportionately from operating companies and patent assertion entities, and states plainly that individual inventors have a harder time selling their assets. On an asset basis, operating companies supplied 72% of the sold assets.

That is not because inventors' ideas are worse. It is structural:

• Buyers do not browse. Corporate IP teams managing thousands of patents buy through processes, trusted sources and brokers they already know. A listing on a marketplace, however good, is rarely where they look.
• The market is opaque. One broker describes it as heavily opaque and difficult to navigate, in part because buyers acquire through anonymous LLCs and reporting on IP deals is thin. Without connections, a seller cannot even see who the buyers are.
• Single assets carry more doubt. An operating company selling a family of related patents is selling something a buyer can model. An inventor selling one patent, often with no proof anyone uses it, is asking the buyer to do all the work.

None of this makes an inventor's patent unsellable. It means the default path, file a listing and wait, is the one with the worst odds. If you are still choosing a route, start with the best way to sell a patent.

Five things that measurably move the odds

This is where the data turns useful. Each of these is backed by market numbers, and each is something the seller controls.

1. Give a price

It sounds too simple to matter. It matters a lot. The 2021 IAM analysis found that over ten years of data, packages with pricing guidance sold 73% more often than packages without, 30% against 18%, and they sold faster. Yet in 2021 fewer than half of packages came with any pricing guidance at all.

A buyer who has to guess your price will usually assume you are unrealistic and move on to the next package in the queue. Give a number, or at least a range.

2. Prove someone uses it

Evidence of use, typically a claim chart mapping your claims onto a product on the market, is the single strongest signal a buyer gets. According to the same broker's summary of the Richardson Oliver data, packages offered with evidence of use were 82% more likely to be purchased, at an estimated 28% price premium. Only about a third of packages carried it.

This is also why we do not send anything to a buyer that has not passed our own validation gate: claim reconstruction, detectability analysis and a validity probability before a buyer sees a teaser. A buyer who distrusts the listing never opens the data room.

3. Keep the package tight

Buyers generally favour smaller packages. In the 2019 data, single assets and packages of two to five assets sold at around 12%, better than their larger counterparts once the counting method is accounted for. Do not pad a strong patent with weak ones to look bigger. A buyer pricing the package will discount for the filler.

4. Be ready for the first eight months

Because most sales happen early, preparation has to be finished before launch, not during it. That means clean chain of title, maintenance fees paid, a plain-language claim summary, and answers to the diligence list that is coming anyway. Every gap is a discount the buyer applies for you.

5. Get represented

The Richardson Oliver data is blunt: brokers were involved in 87% of sold packages. If a package sold, there is an 87% chance a broker represented it. Brokers are not magic, and their own close rates vary widely. A broker closing 30% of packages is described in that broker summary as particularly successful. But representation is clearly the norm among the deals that close.

If you want to know what that relationship involves, read what a patent broker does. The economics matter too. Our brokerage is success-based: the seller pays commission only on completion, with a minimum fee of EUR 5,000 per completed transaction. No sale, no fee. Whatever broker you choose, be wary of models that charge you up front while the odds above apply.

When not selling is the right answer

An honest answer to "how hard is it to sell a patent" includes the cases where you should not try.

• The patent reads on nothing anyone sells. Without evidence of use, a sale is a long shot. Licensing or strategic abandonment may cost you less than years of renewal fees.
• You still need the rights. If your own product depends on it, a licence may beat a sale. We walk through that choice in license or sell.
• It is still an application. A pending application can be sold, but the buyer is pricing in uncertainty about what will be granted. See selling a pending application before you decide the timing.

And if the patent is strong, used and cleanly owned, the question is not whether it can sell but who will buy it. That is a narrower list than most sellers expect, which is the subject of who buys patents.

FAQ

What percentage of patents actually sell?

Nobody tracks every patent, but the measured brokered market gives a benchmark: over several years, roughly 20% to 30% of listed packages sell, depending on the listing year. Packages listed without pricing or evidence of use do worse.

Can I sell a patent without a broker?

Yes. A direct sale works well when you already know the buyer, for example a company you have talked to about the technology. But most packages that sell on the open market had a broker, because finding a buyer is the hard part.

Is a pending application harder to sell than a granted patent?

Usually. A buyer of a pending application cannot know the final claim scope, so they either pay less or pass. Independent validation of the claims before going to market reduces, but does not remove, that uncertainty.

Does an old listing still have a chance?

Yes. There is a long tail of sales years after listing. Older packages tend to sell at larger discounts to asking, so a refresh with a realistic price and better evidence is often worth more than waiting.