use of network patent analysis (npa) for advanced analysis of patent data creating business...
TRANSCRIPT
Use of Network Patent Analysis (NPA) for advanced analysis of patent data
Creating business advantage from patent insight
Mike Lloyd, Doris Spielthenner,
and George Mokdsi
5th June 2012
1. Introduction to NPA
Patents are becoming big business…
But which patents? It is self-evident that patents have a range of values..
0
20
40
60
80
100
0 50 100 150 200
Relative patent value (highest ranked patent = 100)
Patent ranking(1 = highest ranked patent)
Relative patent value distribution
Hybrid car patents (58,000 patents)Alzheimer's treatment patents (2152 patents)Smartphone patents (7093 patents)
NPA is based on patent citations
Patent citations
Each patent citation provides two powerful and valuable insights
1) Both patent applicants thought that the subject matter they filed their patents in was important and valuable enough to invest in a patent filing (whether the later applicant knew about the earlier patent or not)
2) Either the examiner or applicant for the later patent thought that the earlier patent was similar enough to disclose a similar feature. In other words, human intelligence has been used to make connections between patents.
Most patents have both forward and backward citations
Subject patent
Reverse or backward citations
Earlier patents considered to be similar
Forward citations
Later patents considered to be similar
Citations in turn lead to other citations..
US 54870691996
US 56129481997
US 56129481997
US 56129481997
And 79 other patents..
US 46792271984
US 46303141986
US 36050191971
US 48887671989
US 50955351992
US 51955761993
US 5283780,1994
US xxxx199x
US 56129481997
US 56129481997
Other citations
US 56129481997
US 56129481997
Other citations
But the sheer quantity of citation data can quickly overwhelm ….
NPA uses algorithms to identify the most connected patents, and then cluster and rank patents within these networks:
Image taken from a study of 9,000 smartphone patents
Cluster of related patents
And by doing so, summarising and learning from the collective wisdom* encapsulated in the whole of the citation data in the area you are looking at
* See Surowiecki, James: “The Wisdom of Crowds: Why the Many Are Smarter Than the Few and How Collective Wisdom Shapes Business, Economies, Societies and Nations”, 2004.
Each cluster of patent filings shows an area of high popularity
• Blue – client• Red – competitor• Bigger dot = more influential patent/inventor• Thicker line = stronger relationship• Adjacent patents = more similar patents
Results are taken from an NPA analysis of 250,000 engineering patents
Each citation has a direction, which can provide information about technology development, or ‘knowledge flow’ between competitors
Does NPA work?NPA rankings gave better predictions of Alzheimers drug trial patents than other rankings based on any of:- Forward, backward or total citation count- Number of family members
DrugPatent
protecting drug
NPA rankin
gCount Ranking based on number of (48,000
patents)
Forward citations
Forward citations per year
Backward citations
Total citations
INPADOC family
members
Forward citations
Forward citations per year
Backward citations
Total citations
INPADOC family
members
Bapineuzumab (Phase III
Alzheimers drug)
US7189819 1 14 1.6 304 318 395 ~5100 4942 93 95 308
Solanezumab (Phase III
Alzheimers drug)
US7195761 14 15 1.9 35 50 52 ~4800 3949 ~2750 ~3170 ~4400
Phase II trial Alzheimer's
drugUS7xxxxxx
~around 200th
18 0 0 18 ~30 ~37,000 ~21,400 ~4750 ~9830 ~12,000
Weighted forward citation arrows have correlated to successful litigation
Results are taken from an NPA analysis of 72,000 hybrid car patents
2. NPA in practice
Patent portfolio strength can be quantified, on a cluster by cluster basis
Up and coming patents can determined soon after filing, allowing analytical prediction of potential technology trends
Filing year 2005 2006 2007 2008 2009 2010
Leading smartphone of that year, (CNET, other sources)
Leading NPA patents in smartphone study filed in that year
US20060161871, Proximity detector in handheld device, Apple, NPA ranking = 2
US7764274, Capacitive sensing arrangement, Apple, NPA ranking = 4th equal,
US7812828, Ellipse fitting for multi-touch surfaces, Apple, NPA ranking = 9
US7479949, Touch screen device, method, and graphical user interface for determining commands by applying heuristics, Apple, NPA ranking = 146
US20090244031, Contact Tracking and Identification Module for Touch Sensing, Apple, NPA ranking = 424
USD628546. Mobile Phone, Samsung, NPA ranking = 624
Patent thickets can be quantified
Mining project
Mechanical engineering project
ICT project
Litigated smartphone technologies
Alzheimer's treatments
0 5 10 15 20 25
Largest cluster
2nd largest cluster
3rd largest cluster
Cluster densities: average relationships strength within clusters (patents < 20 years)
Relative technology turnover can be quantified
Mining industry project (n=4100)
Mechanical engineering project (n=7450)
Food industry project (n=24,000)
Smartphone - Touchscreens (n=8900)
Smartphone - Mobile data transmission (n=6500)
Smartphone - Mobile data access (n=12,800)
Alzheimer's drugs (n=17,800)
ICT project (n=21,100)
0 5 10 15 20 25 30 35
Average age difference between patents and their backward ci-tations in the same field (years)
'Patent turnover' - 9 out of 10 backward citations are younger than this (years)
Number of years between the publication dates of patents and their backward citations
NPA in software form
• Client studies currently undertaken using consulting model
• AmberScope’ is now in beta testing– Will provide interactive online view of patent landscape
around a nominated patent (‘houses in your street’)– Ask me during the seminar for a demonstration
• ‘AmberMap’ is being developed– Will provide an interactive online view of the patent
landscape in a technology (‘houses and suburbs in your city’)
– Similar to what has been presented today, but with new capabilities
• ‘Further products in pipeline• See www.ambercite.com for more information and
updates
www.griffithhack.com.au
Thank you for your time… any questions?
AmberScope (currently proof of concept)
Does NPA work? (1) Top ranked patent in 2011 NPA smartphone study independently favoured by leading smartphone patent wars commentator
• Florian Mueller, 14 January 2012, “fosspatents.blogspot.com” :
“…but it isn't nearly as essential to Apple's litigation strategy as ..: U.S. Patent No. 7,663,607 on a "multipoint touchscreen". The '607 patent is the broadest touchscreen-related hardware patent Apple has, and if the courts interpreted it as broadly as Apple would like them to, ..would be extremely hard to work around.”
3. Using NPA for advanced analysis of patent data
NPA is a unique means of clustering patents…
• Much more precise and reliable than IPC codes
• Avoids issues due to inconsistent choice of keywords, or keywords covering broad technical concepts
• Groups inventions together, not just technologies
NPA is a unique means of ranking patents
• Uses much more data than a simple forward citation count, in particular focusing on the broader picture
• Provides an indication of underlying patent quality that is more based on peer review than self-assessment (such as litigation measures or the size of the family)
• Compares patents against their closest competitors, not a broad field
Contents1. Why NPA? An introduction to Network Patent Analysis
(NPA)2. How NPA differs from other analysis techniques3. Using NPA for advanced analysis of patent data4. Summary and conclusions
Or even just to provide a unique perspective on individual patents
NPA data can provide a unique perspective on the likelihood of successful patent assertion
Lines showing citation linkages between cluster of Apple patents and asserted
Motorola patent
Cluster of Apple touch-screen smartphone patents
Too far?
Motorola touchscreen patent US 6,246,862, asserted against Apple