How does it work? Taking citation analysis to the next level

How did we get here?

Patent analytics
problems today

Unreliable patent to technology mapping
Using keywords and CPC codes for technology mapping creates inaccurate, incomplete, and inconsistent results. Yet, they are still used today by almost every patent analytics tool. Because of this, it is very difficult to get reliable insights within a reasonable amount of time.
No reliable patent scoring
Many patents are not very important, so without a reliable quality score, we still know very little about the competition. Unfortunately, currently available scores are based on flimsy evidence, do not measure quality well, and are thus not used for decision making often.
Bad intelligence leads to poor decision making
Without a reliable map of the environment, it is very difficult to arrive at your destination. Bad intelligence leads to poor decision making, limited value creation, and in the worst case, disaster.

What makes Focus different

Making patent analytics strategic

Tech mapping

Use your domain knowledge to teach Focus' AI what to look for by example. You can train Focus to learn your internal taxonomy, or any other technological category, and let it mine through any scope of patents for matches.

Scoring

Focus' quality scoring algorithm accurately predicts real-life events such as when a patent will lapse and whether it is likely to be used in litigation. Allowing you to accurately estimate risk, value, impact, etc, in any portfolio or tech domain.

Strategic

Communicate the strategic value of IP by condensing extremely large volumes of data into board level insights that go beyond patents alone.

Time

Very little manual effort is required to operate Focus' algorithms and arrive at the insights you need. Leaving you with more time for strategy.

Get the research

Beyond counting citations

Most patent scoring methods are based on citing rates one way or another, but patent quality depends on many other factors as well. Claim broadness and strength, family size, prosecution history, etc, all tell you something about the likely quality of a patent. Focus developed over 30 proprietary metrics that measure different aspects of patent quality. These metrics are calculated for every patent published since 1995.

Deep learning for scoring

For many of the patents that were published since 1995, we know when they lapsed and whether they were ever litigated. This gives us the opportunity to learn what set of characteristics leads to what outcome. Rather than deciding the weight for every variable by ourselves, Focus uses neural networks to model what combination of characteristics makes a patent have a certain outcome. Neural networks are perfect for a task like this, because they can model very complex relationships. The resulting model is then applied to make predictions about patents for which the outcome is yet unknown.

Schedule demo

How do we compare to the competition?

Focus

The Rest

Learns your taxonomy

Tech mapping

2-5X more patents avg. 97% accurate

Incomplete
and noisy

Reliable scoring

Nope

Easily obtain strategic insights

Save time and money

Nope

Common Questions

Let's put your mind at ease.

Is Focus AI here to eventually replace me?

Absolutely not, Focus' deep learning models are essentially pattern finding machines that can derive meaningful insights from very large volumes of data. What they cannot do however, is come up with a plan to deal with those insights. Computers crunch numbers, people set strategies. Think of our AI as a supercomputer assistant that can be used to leverage your industry expertise.

Why deep learning?

Deep learning models are at their best when there is lots of training data available. Patent data is perfect for deep learning because there is so much of it, it is highly structured, and generally of high quality. The combination of these factors make that deep learning models perform so well with patent data.

Can patent importance even be measured?

Patent importance is hard to define in absolute terms but can be measured by proxy. Most people would agree that on average, a patent that was maintained for 20 years is more important than a patent that only lasted for 4. Similarly, a patent being litigated is a clear value signal. Focus' deep learning models learn what characteristics make a patent likely to lapse within a certain number of years, or to be used in litigation.

Where do you get your data from?

Our patent data is provided by IFIclaims, who provide over 125 million patent records from over 98 countries. IFIclaims also provides data to services such as Google Patents, Questel's Orbit, and many more.

How does it actually work?

How did we get here?

Patent analytics
problems today

Unreliable patent to technology mapping

No reliable patent scoring

Bad intelligence leads to poor decision making

What makes Focus different

Making patent analytics strategic

Tech mapping

Scoring

Strategic

Time

Beyond counting citations

Deep learning for scoring

How do we compare to the competition?

Focus

The Rest

Common Questions

Let's put your mind at ease.

Is Focus AI here to eventually replace me?

Why deep learning?

Can patent importance even be measured?

Where do you get your data from?

Get impressed

Plan a personal meeting.

How does it actually work?

How did we get here?

Patent analyticsproblems today

Unreliable patent to technology mapping

No reliable patent scoring

Bad intelligence leads to poor decision making

What makes Focus different

Making patent analytics strategic

Tech mapping

Scoring

Strategic

Time

Beyond counting citations

Deep learning for scoring

How do we compare to the competition?

Focus

The Rest

Common Questions

Let's put your mind at ease.

Is Focus AI here to eventually replace me?

Why deep learning?

Can patent importance even be measured?

Where do you get your data from?

Get impressed

Plan a personal meeting.

Patent analytics
problems today