Everything was ready:
- The migration plan was ready
- The internal and external communications had been published
- The contract was almost signed
- The change management stages had been clearly identified and planned (presentation in offices)
- Pilot users and others were looking forward to fully using the Google environment
- Discussions on the security of the solution had already taken place.

During the latest discussions on the subject between Frédéric (CEO), Virginie (CFO) and I (CIO), we reopened the debate on the confidence that we could place in an outsourced messaging system. Although we reopened it rather late, it is never too late as long as the contract hasn’t been signed. In terms of security and robustness, there was no doubt as to the toughness of Google’s solution systems. As I mentioned in my article on migration, if we leave the storing volume aside (25GB / user), the costs are the same. Our problem pertained to the level of confidence that we could place in the system.

As far as outsourced solutions are concerned, we must be neither paranoid, nor naive. We simply must set our standards according to our needs. As long as a system is capable of indexing email content and, based on this information, it can offer commercials (free version), what prevents the system from reacting to sensitive information in a completely transparent manner?

We therefore changed our minds at the last minute due to the risks that this choice could entail for our customers with whom we sign confidentiality agreements, we work on R&D subjects etc. We didn’t want to ask our customers what they thought about outsourcing our messaging server. When you ask someone about potential fears, their first answer will be that you need to find a method of protection against those fears. We changed our minds despite also knowing that we were losing the considerable gains in productivity that this complete ecosystem could offer us.

There are several people (infrastructure department, pilot users, users of personal Google services) who might be disappointed in our decision, as they might feel that we are abandoning a promised land, but this choice was necessary considering the risks that we were taking. Users need to know that, although we won’t reach the same level of service as Google overnight, we are already deploying a plan in order to come close to that level. As far as the infrastructure is concerned, we are now planning on providing advanced training sessions on our messaging server for strengthening your skills in order to improve our messaging infrastructure (AS, AV, Cluster etc.).

The reversibility plan initially designed for pilot users was launched immediately after we made our decision. Now, we no longer have any Google applications (if I were being paranoid, I would wonder whether they are truly deleted). I feel a little sorry for Antoine (commercial department) who had done a good job with this proposition through sustained monitoring; it’s a pity. But these are also the rules of the game.

Generally speaking, should all ISDs refuse to use Google’s business services? Answers to this question can be very different. The ISD should not make its choices by using fear as an excuse. But, in order to make sure that the ISD is in line with company strategy, the objectives of the company need to be understood. As a nearshore/offshore IT services company, we need to build customer confidence. By integrating Google into our customer relations, there would have been a doubt as to our complete impartiality as there is a real convergence with Google (IT innovation, software publisher etc.). I know publishers, at least one retirement insurance company, universities, SMEs and many others which use Google services. They felt that their level of confidence was sufficient. But in the end, our choice was made between an outsourced solution and an internal solution in order to ensure a higher level of confidentiality in our customer communication.

[Episode 07] Migrating the instant messaging service to Google