Way back in 2013 I wrote an article in FX-MM magazine (no longer operating) about the ‘false economy of the million dollar spreadsheet’.  The article was inspired by the panic that happened across the financial markets in the wake of the financial crisis.  Banks poured lots of resources (and expensive legal review) into pulling out information from contracts and documents, to answer whichever fire-drill was happening that week, be it negative rates, ratings downgrades, close-out and termination provisions or simply collateral eligibility and ownership.

The gist of this was that at that time there were a lot of very expensive spreadsheets floating around.  Point-in-time answers to point-in-time questions.  Rolling the clock forward 7 years – the landscape has changed a bit but in many cases it seems that the throw-away spreadsheet has been replaced by a throw-away (equally expensive) contract analysis platform output.

Project versus strategic thinking

All contract analysis platforms are not created equal.  There is variety in the clever ‘artificial intelligence’, machine learning or rules based data extraction.  There are varying user interfaces or pretty graphs and reports.  At the heart though, the purpose of these tools is to process a set of documents and allow the end user to answer a set of questions on them.  To some extent they all rely on human inputs and reviews however clever they claim to be.

The problem is, that they are designed and used for projects.  For ease of use (and nice sales demos) these tools are often simplistic:

  • they work at the single document or scan level
  • they extract entity information from the contracts directly, ignoring the myriad ways that the same legal entity is written down
  • they ignore the grouping of amendments that may dramatically alter the meaning of the contract data
  • they work purely on the text, rather than showing the reviewer scanned image of the printed page, meaning the reviewer misses the handwritten changes
  • they are not designed to handle structured data
  • they have little or no workflow or audit trail and are not designed to manage business as usual contract changes as they come through

These limitations in the design mean that when the question has been answered the project is complete.  The data is exported (quite often as a spreadsheet still) and at that very moment everyone is happy, confident in the answer.

A few months later the data is either obsolete (new contracts or amendments have come through), or mistakes found, or there is an entirely new question to be answered on some or all of the contracts.  The process starts again as a new project.

This is a very expensive way of managing contract data.

I wonder just how many times a contract passes through one of these systems to answer a question.   Take your typical master agreement.  It will have been reviewed for the financial crisis, but then subsequently perhaps as part of margin reform, record keeping, Brexit, various market and ratings events and now LIBOR benchmark reform.   

I bet that each time, the contract has been loaded up, entities extracted, work performed to determine the amendments and overlay them in the output and a number of the same data points re-captured.  Given the imperfections of automated capture, the manually reviewed (and corrected) data points have also now been lost.

Contract data management versus contract analysis

At some stage it is just economically far more compelling to process your contracts once and to do it properly.  De-duplicate your scans, organise your contracts in the way that you conduct business (probably by counterparty), group the originals and amendments together and relate all contracts for an entity to simplify cross referencing and reporting.   I firmly believe that contract data management is different to contract analysis and that the software is different for both.  

There is fractionally more up-front work to do to set up a contract data management platform.  Going through the rigour of tying contracts down to actual entity reference data, rather than simple textual names.  Weeding out the duplicates, unexecuted copies and sorting the scans into chronological contract families.  Identifying your inactive contracts.  But there are savings to be had from doing this, even on the first project, let alone any subsequent re-visits.

We operate contract data management software.  

Our clients enjoy the benefits of their investment by being able to answer questions on their contracts, quickly and as the need arises without a new project.  They process their amendments  and new contracts as a matter of business operation and not as a panic during a project.  This requires minimal effort as only changes are processed.   They never need to process the same scan multiple times.  Over time they improve data quality in the platform.  They can also use the system for contracts that contain operational data that must be processed in other banking systems.

If you are tired of throwing away value created in answering contract analysis questions please contact us to find out more about contract data management for your business.