Inprotech: Data Quality

Data quality is always an issue for patent attorney firms.

Many years ago we developed functionality to report upon potential data exceptions within the database. At a recent client meeting we were talking about auditing data and this report was raised, once more a good deal of people in the room were unaware of its existence.

It's called the Property Data Exceptions report and it is available from the Reports menu of the case summary window, the one that you get that lists all the cases selected and allows you to tick cases for reporting and processing. The menu item sort of sneaks in there on this window and you could be forgiven for not noticing it because it is not available on the Reports menu on the first selection window. Unless you look closely it is not so obvious that it is now available on the summary window.

The report was written for CPA renewal clients in attempt to be able to get the data as clean as possible before sending it electronically to CPA via the interfacing programs. As such, it has a bit of a renewals focus but there is some utility for the audit of data in general. It runs over the selection of ticked cases in the summary window and can be selected to run over all the cases or only those with a case type of Property.

Unfortunately there is not a lot of documentation about what the report does (in fact, absolutely none) so the following is what each of the options does to the best of my memory:

1. Agent Details - foreign cases where no agent is recorded against the case.
2. Application Details - the case has an application number but no application date or vice versa.
3. Registration Details - the case has a registration number but no registration date or vice versa.
4. Publication Details - the case has an publication number but no publication date or vice versa.
5. Priority Details - the case has a Basis that is claiming convention but has no earliest priority date.
6. Entity Size - the country requires definition of whether small or large entity parameters are to be used to calculate fees (eg. US and Canada) but this hasn't been recorded against the case. This is determined by an attribute recorded against the country via the Country program.
7. Number of Claims – the country requires the number of claims to be able to calculate the renewal fee, eg. Japan, The Philippines, but the number of claims hasn’t been recorded against the case. This is determined by an attribute recorded against the country via the Country program.
8. Number of Designs – the country requires the number of designs to be able to calculate the renewal fee, eg. Germany, but the number of designs hasn’t been recorded against the case. This is determined by an attribute recorded against the country via the Country program.
9. Date Sequence – makes sure that the dates recorded against the case meet the validation requirements as defined by the Dates Logic rules for the applicable criteria. This is the area where you might define, for example, that the Registration Date needs to be greater that the Application Date. This validation rule will be applied at the time of data entry however any previous loaded cases can be audited overall.
10. Official Number Format – applies any number formatting rules to official numbers entered and displays an error when the validation fails. The formats need to be set up against the number types for the country via the Country program.

Upon reflection, I can think of a number of extensions to the approach to add a lot more value but it is there right now and it is free to use!
One idea could be to use it if you are looking at giving a client access to their portfolio over the web via the Client WorkBench. Bring all their cases up in a query and run the report over it to see what exceptions are highlighted. This way you have an idea of the data quality that you are presenting to the client in a quick report and therefore assess how you want to handle the situation.
Another could be to run the report for cases in a takeover situation after they have been loaded.
Food for thought.