Why Patent Work So Often Comes Too Late

Introduction

I have seen three kinds of companies trip at the same place.

First: the product is about to launch; a search is run; the core structure may hit someone else’s still-valid patent. Second: goods are already on the vessel before anyone notices that the target market has almost no patent coverage—and many filing windows that could have been planned earlier are already gone. Third: dozens of patents are filed every year and maintenance fees are paid without complaint—yet when someone asks which ones would actually hold up in a dispute, no one in the company has an answer.

These three companies do not look alike. Different industries, different sizes, different attitudes toward IP. Some take it seriously: a dedicated team, a real budget, professional counsel. And they are still late.

What became clearer over time is this: the problem is often not only attitude.

1. Three Common Explanations That Do Not Hold

Why so late? The industry usually offers three attributions.

First, blame R&D for not caring. Yet some R&D directors care a great deal and tell teams to “report every innovation”—and they are still late. What they can manage is often only “file after development is done,” not “look at competitors before development starts.”

Second, blame the service provider. Some companies have switched three or four firms, including foreign firms—and the problem remains.

Third, blame the boss for not understanding patents. Yet some bosses come from a technical background, know patents well, and spare no budget—and still find trouble the night before launch.

Three explanations point to three different people. When the same problem repeats across different people, the cause is unlikely to sit in any one of them alone.

2. Draw the Process—and the Gap Appears

Look at where patent work sits on many companies’ timelines.

Project kickoff → review → design → sampling → validation → freeze → mass production → launch. Along that line, patent work often lands late—sometimes after the product is already frozen.

A common chain looks like this:

R&D complete → invention disclosure → instruct counsel → draft application → deliver

Patent work often begins only after R&D is finished. By then, much has already happened: the technical solution is locked and hard to change; the product route is set and market strategy is hard to move; the cost structure is fixed and changes are expensive; the launch date is booked and the calendar resists slip.

In other words, when patent work starts, many adjustments that would once have been cheap have become hard.

That does not mean patent work at this stage is useless. But if the task is still only “get this solution filed,” it easily becomes paperwork that records what was done—not judgment that helps decide what should be done.

3. The Hard Part: Every Step Looks Right

Ask anyone on that chain and they will say their step is fine.

“You can only file once the technology exists.”—True.

“You need a solution before you can draft.”—True.

“Once the papers are delivered, the engagement is done.”—Also true.

Every step is reasonable. That is why the chain is so stubborn: no one sees a problem, and no one clearly owns it. Taken together, it becomes a chain where patents can only record after the fact.

It resembles a familiar management pattern—local optima: each step is sensible alone; the whole is not. It also reinforces itself. Because patents sit at the last step, early R&D does not think about them; because early R&D does not think about them, patents can only appear at the last step. One loop, and the company locks itself in.

4. Why the “Last Step” Position Is Especially Costly

To see why, return to what a patent is.

The core of a patent right is exclusivity. Its value depends largely on two things: how much ground you have claimed, and where others’ claims sit relative to yours. The second point is often overlooked.

If search only happens after R&D is finished, you may learn that the solution carries risk. You can still respond—but molds may already be cut, testing done, the supply chain ready, and change becomes expensive.

If the same search happens during design? The same risk finding may mean changing a connection method, a material mix, or even a technical route—usually at far lower cost.

What shapes the outcome is not only whether you searched or how skilled the searcher was, but where in the process you searched. Position largely decides whether you are a participant or someone being informed after the fact. Participants can still shape the result; the informed mostly face a result that has already formed.

5. Same Facts, Different Timing—Very Different Bills

Abstract talk is less clear than three common scenes.

[Comparison] Found at the end of the process vs found early

Structure may fall within someone else’s patent

• Found late: mold changes, extra certification, delayed shipment—or licensing and other remedies

• Found early: change the connection or structure; adjustment cost is usually lower

Product will be exported, but no filing plan was set early

• Found late: some filing windows are already gone; only residual options remain

• Found early: prioritize markets at kickoff and sequence overseas filings on schedule

A key innovation appears

• Found late: reconstruction after the fact; easier to miss

• Found early: identify and protect while the solution is still forming

All three say the same thing: the same information, received at different times, can be worth very different amounts. Critical risks often do not appear only because you searched late. What changes is when you get the information. Timing largely decides whether that information supports decisions—or mainly supports remediation.

6. Better Node Design: Move Judgment, Don’t Just Add Steps

How should the work nodes be designed? Many people hear “intervene earlier” and assume more paperwork for R&D. It is not simply adding steps. It is moving some of the judgment that used to sit at the end further forward.

In practice, four positions matter:

• At project kickoff: break down the technology, confirm target markets and competitors—answer “where are we going?”

• During design: screen risk and spot design-around options—answer “how do we route around it?”

• As results form: dig for patentable matter and plan coverage in parallel—answer “what ground do we claim?”

• After product freeze: run a focused FTO on the final solution—answer “can we launch?”

The fourth position deserves care. Risk screening during development helps R&D decide; FTO after freeze is a formal gate on the final product. They are different jobs. Early screening does not replace later FTO. The real gap is usually in the first three positions.

7. A Clarification: The Issue Is Not Per-Case Billing—It Is the Node

At this point some colleagues may hear criticism of per-case billing. That is not the point.

Per-case billing is not wrong. One instruction, one delivery—easy to account for, deliver and compare. That model has long had efficiency and value. The problem is not “per case,” but that the service boundary is easily defined as “the document.”

And the problem is often not only people. Many patent attorneys are conscientious. Sometimes they see a problem as soon as they open the disclosure—but by then it is already late. What is draining is not missing the issue; it is seeing it when little room to adjust remains.

So the real issue is how work nodes are designed. You cannot fix “intervened too late” by “writing a better application.” Those are different layers. The document is the result; the node is the cause upstream.

8. What Companies Can Do Without Replacing Anyone

The advice is not to switch counsel the moment this pattern appears. That may not solve it.

Start smaller: move the node one step earlier. Ask three questions:

• In the past year, has anyone who can actually make patent judgments sat in any product kickoff meeting?

• Do invention disclosures arrive from R&D on their own, or does someone have to chase them every time?

• Of the dozens of patents on the books, how many were decided while the solution could still be changed?

After those three answers, where patent work sits in your company is usually clear. Then take one minimal step: at the next kickoff, bring in someone who can make patent judgments—to listen. No long report, no new heavy process. Let them hear who the product will be sold to, who the competitors are, which technical direction matters, and which options are backups.

The cost of that step is usually low. Moving the node one square forward can be the difference between a small redesign later and a major rework.

Conclusion

Return to the opening question: why does patent work so often come too late?

Late is often less about attitude than about position. It is not that no one cares, no one owns it, or that skill is missing. It is that the work has been placed where it can barely influence the outcome. At that point, even excellent counsel has limited room to adjust.

What you actually need—knowing what to change while change is still possible—does not happen at that position. Time does not run backward. The next node can still move forward. Moving one step usually costs little. Leaving it where it is often costs more later.