In the last post in this series, we introduced the topic of often costly – and infrequently useful – patent watches. We highlighted the three significant use cases for alerts – for opposition purposes, current awareness, and continuously-updated competitive intelligence and landscapes. Beyond that, there are many more, such as identifying and monitoring new and interesting developments in key technology areas for acquisition, licensing or partnership, keeping abreast of new trends and potentially-disruptive technologies, etc. When done right, alerts are valuable for all these, and the results can be ‘gold dust.’ Sadly, there are also lots of pitfalls that we see time and again in practice.
This post aims to break down the patent monitoring process into its key steps to answer: what makes the perfect alert?
The key to a successful alerting process depends on its genetic makeup. We can simplify this by looking at the three elements we’ve discussed before, and that apply to any alert and use case you can think of. They are recall, precision and insights, and together they form what we call the Alert Quality Index (AQI). Each are interlinked and essential in the process.
Recall refers to the number of relevant patents or non-patent literature sources identified through the alert process. Ideally, you want your alerts to capture everything that is relevant and miss nothing important in the compromise of limiting the number of results returned. If you have a low level of recall, the insights that can be derived are of a lower quality as not all critical documents will be identified. This in turn results in annoying your stakeholders, as the missing key documents inevitably lead to questions over the value of the alerting.
This is directly linked to the second element, precision, that pertains to the accuracy of the search strategies used. With any search string, you’ll always have a degree of unwanted ‘noise’ and the precision element is the acceptable level of ‘noise’ versus the actual ‘noise’ returned by your strategies. High precision needs to be balanced with the desired recall: this ensures key documents aren’t missed, while keeping their number manageable for reviewers. This balance is vital, but can be hard to achieve: we often see companies fall to one side or the other, inevitably leading to low internal compliance or limiting the value of insights derived.
Insights are the final element in the genetic makeup of an alert. Once you have high recall, with the desired level of precision, the information captured can be used to derive insights above and beyond the initial use case of the alert. It can give companies the list of patents for post-grant opposition, while also catering for updated competitive intelligence and technology landscapes, essentially becoming a searchable central repository for the company’s technology universe. This is often overlooked but is a key to deriving economic value, as we’ll discuss later.
These three interlinked elements form what we call the Alert Quality Index (AQI):
AQI = (RA/RE + PA/PE + IA/IE)/3;
with RA,E = Recall actual and expected, PA,E = Precision actual and expected and IA,E = Insights actual and expected.
Using the formula and the list of use cases, we can now do a systematic analysis of how an alert should be set up to deliver value:
|Use Case||Main Deliverable||Impact of Recall||Impact of Precision||Other factors impacting quality|
|Identification of cases for post-grant opposition cases||A list of patents for post-grant opposition||High recall is desired, as all critical patents should be identified through this process||High precision is desired as otherwise workload of internal reviewers is too high and compliance low||N/A|
|Regularly-updated patent landscape and/or competitive intelligence||A patent landscape||High recall is desired, as otherwise the insights from the landscape are not correct/ precise||High precision is desired as otherwise the insights from the landscape are not correct/ precise||Customized, deep indexing will further improve the quality of insights – an appropriate tool for visualisation|
|Current awareness||A repository of patents/ NPL of interest||Medium recall is acceptable, as the aim of this alert is to be informed on recent developments||Medium precision is acceptable. as the aim of this alert is to be informed on recent developments|
So, we now have the genetic makeup of a perfect alert with the desired level of recall and precision that will allow us to gain the insights we need. ”This is great!” I hear you say, “we’ve done it!” I hear you cry. If only it were that easy. Capturing the optimal data set is only the first step on the journey.
In part 3 of this series on IPR&D alerts, we’ll look at how to unlock the true power of your patent watch efforts. We’ll cover how organizations can effectively leverage captured data, and how companies can unlock economic value from their alerting investment and effort. You have come this far on the journey and are now one step closer to reaching the Nirvana of all alerting processes! We hope to see you there!