Lux Magi
Productivity Through Experience

Example Investment Banking Engagement: Rule Based Accounting


This is a precis of a nine month assignment we undertook for a top-tier American investment bank. It is designed to give prospective clients an example of the style of projects in which we are involved.

Project

This project concerned the architecture and development of a stock record and trial balance system for stock loans, FX and repos with real time subledger support. The scope included business event validation, classification and accounting. The key business drivers were:
  • shared business and IT governance of the accounting policy
  • high agility - quick turn around time and low error rate for policy changes
  • ability to explain results after the fact
  • integration within the defined finance business process model

The client had already captured 3000+ accounting rules in spreadsheets and implemented them in rules using a well known BRMS. The resulting implementation had several issues.

Initiation - Being Effective from the Start

Lux Magi was short listed for this engagement by one of our partners on the basis of our specialised experience with rule based accounting systems. Our consultant’s suitability was tested by a single three hour workshop with the client. During this workshop we assessed the problems faced by the client and warned them of other potential architectural pitfalls that had not previously occurred to them. Because of our business specific experience, we were able to give specific, cogent and useful advice right from the outset. On this basis we were selected.
In a follow up workshop, our consultant summarised the key business challenges and made specific recommendations to address each of them:
  • high number of manual exceptions - we proposed changes in the architecture and business process to increase the straight through processing rate
  • performance issues - we proposed several measures to enhance scalability and a specific program of profiling to uncover bottlenecks
  • lack of business visibility of the accounting policy - we recommended a replacement BRMS to allow the client’s own business team to take ownership of accounting business rules
  • lack of a coherent business policy governance procedure and change process - we suggested that we work together to make explicit the client’s own governance process and augment it with certain industry best practices
  • high rule entry error rate - we proposed changes in rule structure to reduce opportunity for human error
  • changes to policy taking too long to implement - we proposed several practical means of improving agility and a fact find process to discover more
  • excessively large and difficult to understand rule set with overly complex rules - we showed how refactoring could benefit clarity and size of the ruleset
  • reliability issues - by adopting a state of the art BRMS and using automated regression testing we showed how reliability could be considerably improved
During the week that followed, we worked with the client to:
  • Produce a short list of measurable metrics (visibility, agility, rule repository size, rule complexity, performance, reliability and straight through processing rate) and target goals
  • Devised a plan to address each goal, including a demonstration of improvement after four weeks
  • Devise a longer term hand off and knowledge transfer back to the client (see our Mutual Development Value)

Delivering Value Quickly

It is always our policy to demonstrate tangible business value (using a proof of concept or prototype) in as short a period as is practical. In this case we were able to implement a prototype accounting component (handling six specific business events) within four weeks. This prototype addressed the most significant 10% of the business scope and showcased:
  • a reuse technique for reducing the accounting rule set size by a factor of three
  • a new rule idiom for reducing the complexity of rules by five times (20% of the original conditions remained covering the same business logic)
  • reduction of rule entry error rate by empowering the rule engine to spot omissions and overlaps in rules where it was previously unable
  • improvement of error rate and business visibility by new rule entry mechanism and use of strong datatypes (rather than strings)
  • agility enhancement by auditing rule behaviour and providing users with change impact assessment reports
  • agility increase by reducing business/IT hand-offs from five to three and allowing the business to change some rules with no IT involvement
  • complete resolution of reliability issues by use of new toolset
  • demonstrated the business value of accounting journals that justify their creation (by traceability to accounting policy)

Productivity Through Experience

Having demonstrated value early, we then set about optimising our productivity. Over the next two months (phase one) we completed:
  • Implementation of complete end-to-end accounting for one third of the business scope, with business sign-off of all output
  • Knowledge transfer to allow rule implementation layer to be handed over to the IT team
  • Knowledge transfer to allow rule base to be handed over to the business team
  • Implementation of all key data entry screens and views
  • Implementation of all key data pathways (database, query helpers, enumeration integration)
  • Integration with host technical architecture with changes to the latter where required
  • Measurement of performance and identification of bottlenecks
  • Improved business rule verbalisations for compliance with official business terminology

In implementing the core of the system in the first 33% of the timescale, we significantly lowered the delivery risk of the project. The rest of the effort was essentially ‘more of the same’ - an opportunity to coach the business teams as they developed rules for themselves. Issues identified by the first phase (reliability, performance) were then addressed.

Knowledge Transfer

During the next three months (phase two) the business were coached as they produced the next third of the business logic. Towards the end of this period they became completely self sufficient, being entirely responsible for 40% of rule change scenarios (those not requiring new technical assets, data, etc) hence satisfying the visibility goal (see above). In addition we achieved:
  • Agility/Process: Instigation of a new process to control collaboration of IT and the business in the case of changes requiring joint work; this reduced the average change time from three weeks to two days.
  • Visibility: business having taken ownership of rules now took ownership of verbalisations (natural language phrases for business concepts) and an accounting glossary
  • Error Rate: the error rate was reduced to less the 3% of its original level
  • Repository Size: was reduced to 35% of the original
  • Rule Complexity: was reduced by a factor of 90% (when measured as number of conditions expressed)
  • Performance: was improved by a factor of 400%
  • Reliability: was improved by a factor of 70% (by a change in repository structure and hardware), but some new issues were discovered
  • Hand/Over: some advanced rule design issues were handed over, albeit later than planned

Conclusions

All of the remaining issues were resolved during phase three and performance was improved by another 100%. The project was delivered on time and on budget. Our consultants produced a full documentation pack of all assets, processes and technical procedures they developed during the project.

Like to Know More?