Extreme Transactionality and Business Metrics


We'll be standing on the shoulders of FileMaker giants, and sitting on the floor with some - such as Todd Geist, Wim Decorte, Corn Walker, Kevin Frank, Ray Cologon, and Mikhail Edoshin - looking at techniques each of them developed and/or championed. We'll be exploring and extending techiques developed by these bright guys, leveraging the FileMaker engine and UI. We'll consider the wisdom of the FileMaker engineers and also the relational database theorists and designers from other environments.

What's the problem?

It usually matters that the debits equal the credits or that the number of items that the database says are on hand matches the quantity on the shelf. And fast reporting matters too. For instance, multi-column income statements with different ways of aggregating data... or inventory trends by product group by location by month. The solution: pre-aggregate the information in a data warehouse table.

Then What?

Depending on the situation, the reports can be produced directly from the aggregated data or extracted into dedicated tables using complex keys and/or multi-predicate relationships.

What do you do with the data? How do you turn it into useful information? Trends? Ratios? KPI's (key performance indicators)?

Techniques

We'll be looking at optimizing performance and managing data integrity with large data sets in multiple tables. We'll explore and discuss creating "data cubes" for managing reporting. The data examples used will be business operations data - both inventory and accounting data - but the methods can be applied to any context where it is essential that data be reliable in multiple tables or multiple locations. It's especially applicable to large multi-user environments, though relevant any time when it is necessary for the whole to equal the sum of the parts... and still be fast.

The techniques can be used in pure FMP environments, or with ESS and mixed FMP/SQL environments.

I think I know something about the topic, but I have many questions and I trust you do too: How many records in a portal? How many portals in a portal? (trick question) Scratch tables or virtual lists? Repeating variables or named variables? Multi-line keys or global keys?

About Me

I've been doing FileMaker development and business systems consulting for 24 years. I have co-developed Genesis Transactions, Genesis Accounting, and Genesis Business Metrics -- which has allowed me to see these problems in the real world time and again and to have bigger and bigger data sets (and hence performance problems) to work with. I recently redesigned a technique for storing data in a cube in our "Metrics Data" table, resulting in a 1000 times performance increase.

I'm a business systems expert first and a FileMaker jock second. I have collaborated with some of the best minds in the business to co-develop solutions. It helps to have smart friends.
Depending on the situation, the reports can be produced directly from the aggregated data or extracted into dedicated tables using complex keys and/or multi-predicate relationships.