Boosting patents won’t solve Canada’s innovation problem

Open this photo in gallery:

Peter Carrescia is a co-founder of Qui Identity, a Toronto-based company focused on decentralized identity solutions. Previously, he led a national technology accelerator and was a managing director at OMERS Ventures.

Bill Gurley, a leading venture capitalist and partner at San Francisco-based Benchmark, calls them “innovation blockers.” Fred Wilson, another top VC and partner at Union Square Ventures, calls them a “tax on innovation.”

The target of their ire? Patents, and specifically software patents. So why, when the smartest money in venture investing believes software patents stifle innovation, is Canada doubling down with programs and funding targeted at startups, initiatives primarily designed to protect intellectual property (IP) through the use of patents?

With the plethora of recently launched government IP programs, and the effusive way our leaders talk about them, one would think there must be strong evidence that software patents are important for the success of a technology business.

But there is scant evidence that is the case. Consider Shopify Inc. SHOP-T, which rose to its dominant position without the use of patents. Not only did the company not have core patents that drove its growth, but also none of the literally thousands of e-commerce patents issued to Amazon.com Inc. AMZN-Q, eBay Inc. EBAY-Q, International Business Machines Corp. IBM-N and countless others prior to Shopify’s rise were able to stop its ascent either.

At the other end of the spectrum, IBM has filed more U.S. patents than any other company every year for 28 consecutive years to 2021, and this includes 3,600 patents in the hot area of artificial intelligence. Despite that, IBM has seen its once-dominant market position deteriorate steadily.

Over the past 23 years, IBM’s market capitalization has declined by almost 50 per cent, and revenue has declined from a high of US$107-billion in 2011 to US$57-billion in 2021. And while IBM generates some IP licensing fees from its patent portfolio, the fees are immaterial and less than 1 per cent of IBM’s revenue.

Clearly the links between innovation, as quantified by patents, and commercial success are weak at best. And these examples are not exceptions. Most successful software startups that have scaled up have a similar story, having succeeded without the use of a relevant patent that protected what made their company successful.

So, if software patents aren’t needed to become successful, or needed to enhance existing success, why are they used? From the perspective of large established companies, patenting can be an effective defensive strategy.

But no one should mistake these efforts as an indicator of innovation that is valuable and needs to be protected. There is far more evidence that large companies, now a bigger target of lawsuits than ever before, use patents almost like insurance. They build patent portfolios to protect themselves with the ability to countersue, cross-license, generate fees from partners and sometimes slow smaller competitors with expensive and often frivolous litigation.

It is problematic, though, to apply big-company patent strategies to earlier-stage companies. First, it is expensive for a startup to patent. In addition to direct legal costs, defocusing engineers and founders away from building software to interface with lawyers and patent agents is hard to justify.

Patent law requires provisional patents be filed prior to, or very near, the time they are disclosed in a shipping product. But the reality of software startups is that they iterate the product with early customers, continuously, and it is very difficult to know which iteration is worth patenting.

Also, with limited resources, early-stage investors want to know the company is focused on finding product-market fit and customer traction, not on a possible patent for something with unknown customer demand.

The act of filing a patent requires public disclosure, as well. But for a startup, keeping the invention a trade secret often gives greater flexibility to experiment with less fear of copycats that might be tipped off and motivated to work around the patent.

Sometimes software patents for early-stage companies can be beneficial, but these situations are exceedingly rare. It can make sense, for example, if interoperability between customers is required, and it is possible to own the patent for the interoperability standard, such as Qualcomm with its CDMA (code-division multiple access) standard. Today, however, standards bodies, which often control these decisions, are so concerned with the risk of a toll on the use of standards that opportunities to own this kind of patent are quite rare.

For all of these reasons and others, almost without exception, successful seed and early-stage VCs will tell their software investees to not bother with patents.

Creating policy that pushes patents regardless of area or company stage and gauges success by counting patents is misguided and, in fact, dangerous to the success of startups. Software startups are almost always far better off if they allocate precious resources to understanding customer needs, building corresponding IP through rapid iteration and then protecting it in ways other than patents.