IT provider changes need a compelling concept
// by Matthias Rensing
IT & Management Consulting, IT Sourcing, Provider Management
Without the comprehensive involvement of the business departments, changing the IT provider is always a risky undertaking
Even if the new IT provider and the customer have already worked together successfully in previous years, the renewed service takeover after a provider intermezzo is still not a piece of cake and can become a lesson. noventum consulting has accompanied such a case in an advisory capacity and the realization on all sides after successful project completion is: the combination of a provider change with extensive system changes must not be overstretched if time pressure is high at the same time. The early involvement of all affected departments and persons in the transition is another important requirement for any future provider change. Despite all the routine and expertise on both sides, the transition can always be challenging. Even intimate knowledge of the IT landscape that is to be taken over does not change this.
The reason for the renewed takeover of the IT services of the customer, a billing service provider with national and international focus, was the routine re-tendering of its IT at the end of the contract period. The in-depth knowledge of the business processes of the customer and its parent company was certainly an advantage for the accepting provider during the tendering process, so that the customer was successfully won back. Having been active in the IT hosting sector for more than 20 years and entrusted with the support of numerous systems of national and international customers, the old/new provider could initially assume that the takeover would go smoothly.
noventum consulting's sourcing consultants were listed as transition experts at the provider from successful projects in the past. In previous projects, noventum had represented customers to the service provider, but now the noventum consultants were to represent a holistic view of the project on his side, taking into account both the customer and the provider perspective.
SAP environment for Europe-wide billing processes
The most important component of the customer's application landscape is a powerful SAP environment, which the customer uses for its accounting runs. With many thousands of acceptance points, the company operates one of the largest sales networks in its industry. Mass billing runs go around (several times a month) between customers, middlemen and retailers in two directions. The smooth functioning of these settlement runs is critical to the business and delays in a run have a significant impact. For this reason, a dedicated SLA has been agreed for the availability and performance of all systems involved in the settlement run, even if the customer is responsible and controls the run itself. The smooth functioning of these services is linked in the SLA to hard contractual clauses and penalties, which underlines the extraordinary importance of the SAP system.
The transition of the IT comprised the following systems in total:
- SAP system
- Windows systems
- Linux systems
- Network components, connections
- Internet access, VPN etc.
In addition to the transition, there were extensive innovations
The sourcing customer took the provider change as a starting point for extensive new installations in the target environment. The experts at the customer and provider agreed that these updates promised valuable improvements. In addition to new hardware, the provider change in the target system also resulted in several new applications, platforms and links, such as the replacement of a Linux environment with an AIX Power environment. Other applications were also updated. The risks of this provider change were therefore not insignificant.
The customer created a structured overview of all application components, starting with the SAP landscape and ending with smaller applications and the remaining data basis. The background to this was that the provider was to provide only the operating system, databases, and data infrastructure for all other applications, with the exception of SAP applications.
The customer's SAP manager had a particularly demanding job to do in migrating his department from Linux to AIX. All data was backed up from a backup to a Linux installation at the provider. The platform change to AIX was then carried out using SAP export/import procedures. This complex procedure was necessary due to the different endianness (byte order) of the x86 and power systems.
The planned system changes, both large and small, were manifold. Bringing the transition forward by one month at the customer's request increased the level of requirements for all those involved.
Customer and provider installed internal and external project managers and consultants for the transition. On the customer side, IT was primarily involved in the planning process, while those responsible for the various application areas at the customer (e.g. process managers) only became active in the steering and planning committees later on.
Successful transition and system change at second attempt
The cutover ran smoothly for the first few hours. Applications were shut down, data was extracted and transported via the migration line to the new data center. The Linux systems for the SAP migration were successfully set up and the export/import procedure started. Several errors occurred, which delayed the export of some systems very much. Despite repeated analyses and attempts, the exports on the cutover weekend behaved differently than during the repeated test runs. On Sunday it was jointly decided to carry out a rollback in parts.
In an ad-hoc session with all application managers of the customer and the system managers of the provider it was determined which environments could already start in the target environment and which were affected by the case back. This plan was executed and production was resumed in the new situation. Some of the old systems still had to be in service for some time.
An in-depth analysis of the cutover revealed that various sources of error together were responsible for the unexpected runtimes of the export/import and another system error. This affected infrastructure capacities as well as application and data logic.
This made it possible to avoid the mistakes in the second attempt and to complete the move within the originally estimated time frame. Marc Buzina, noventum consultant and project manager on the side of the old/new provider is very satisfied with the result despite all adversities. "Dealing with errors and crises also needs to be skilful. We have reacted in an equally professional manner on the side of both provider and customer to the difficulties that arose in data transfer. Today, both are satisfied with the result".
Good crisis management must also be able to say "stop"
Question to the consultant: "What are the consequences and lessons to be learned from a provider change that does not work at the first attempt?
Marc Buzina: "Every transition is special in detail, so you can't have a recipe up your sleeve for every eventuality. But that's not bad as long as there is a clear division of roles. And this is also where the analysis has to start when troubleshooting later on.
- In order to enforce the requirements that one has as a provider, one must take the opportunity to really get all those substantially involved at the table. Escalation channels should also be used in the event of technical differences.
- It is not new and yet it is also important to align test scenarios as closely as possible to reality. Any deviation can become a problem. Sufficient time and capacity must be planned for this at the provider and customer. If necessary, this means carrying out a complete dress rehearsal".
Question: "Should the policy competence for the rules of a transition lie undiminished with the provider?
Marc Buzina: "That's where she lies anyway. The provider controls the transition per se. As a rule, it is not a customer's core competence to control the transfer of its IT from one provider to the next. Nevertheless, not all responsibility can be outsourced. Because this is the case, customers must and should obtain appropriate consulting expertise in-house for this phase. Preferably a good one. “