So, what where the motivators providing the impetus to bring our gift card platform in house?
- We decided to change our credit and debit card processor and couldn't migrate our gift cards to our new processor.
- We knew the time to develop it would only be a couple of weeks of my free labor.
- Internet connectivity is more pervasive and reliable allowing us to bring our gift card platform in house and run as a web service with varied user interface options.
- We were paying around .15 cents USD per transaction.
- It costs us nothing to host it on our web server and it costs us nothing per transaction.
How did we do it?
- I built tables that resembled the data format of our gift card providers database into one of our core RDBMSes.
- I built a back office management suite for managing terminals and terminal ids, card records and transactions.
- I built an authorization engine composed of message parsing and abstraction, rdbms interfacing, message routing and control and logic processing.
- I wrapped the authorization engine in a web service.
- I wrote a front end web interface that communicated with the web service.
- I wrote a VB desktop application that interfaced with a USB magnetic stripe reader and interfaced it with the web service.
The entire system was built in a basic LAMP stack and took about two weeks to build from soup to nuts and we were live. We deployed the VB gift card desk top application to replace the use of the hardware point of sale terminals. The system behaves exactly like the old system except that our staff do not use the hardware point of sale terminals to process gift cards.
So, was it worth it?
Yes of course because of the following.
- We would have been forced to run two hardware point of sale terminals had we not. One for credit and debit card processing. One for gift card processing.
- Training and administration with two pieces of hardware would have had general administrative over head costs.
- We were paying about .15 per transaction to our processor for gift card transactions.
- From launch around mid July 2011 we have processed around 1,258 transactions which would have cost us around $200 in transaction fees.
In summary, this project only made sense from a break-even investment point of view because my labor was free and the cost of additional hardware had we not done this would have amounted to probably somewhere around $10,000 give or take a bit.