Data Conversion Team Lead Best Practices

Jimbo's picture

There are a handful of things that fall outside of the realms of common sense and the obvious. Managing the Extract, Transform and Load (ETL) aspect of an SAP implementation can be a daunting task were it not that there are ways to make the project run more smoothly.

These best practices were developed over the last eight years and come from managers, team leads and conversion specialists who have all been instrumental in successful SAP implementations around the world. What we have learned from our peers and from our mistakes is presented here to help managers putting together a data conversion team and those who have been placed in the team lead position.

Keep the team tight and onsite

Outsourcing the data migration aspects of an SAP implementation to an army of programmers on another continent may be a little less expensive than having a few competent resources on-hand, but the temptation to cut corners in order to save a little has led too many projects to the brink of failure. Experience has shown that a team of two to four competent programmers on-site will be more effective than an army of remote resources every time. This is especially true when the programmers have worked together as a team on previous projects.

Make sure the team lead is a programmer

Programmers make the best managers and trainers of other programmers. Programmers can identify strengths in other team members and distribute tasks accordingly based on abilities. An experienced data conversion specialist can ensure continuity and increase team cohesion by training inexperienced team members in the use of the tools used to perform data conversion tasks.

Distribute tasks in a meaningful way

Besides the generic BASIS and configuration, the first things to be loaded during an implementation are the big three master data objects: customers, vendors and materials. Downstream transactional data cannot be loaded until the dependencies are loaded; there is a flow that can be identified and applied to the distribution of tasks. Assigning transactional data loading tasks to the resource responsible for loading the associated master data eliminates idle time and reduces time wasted searching for missing dependencies. A experienced team lead knows what objects are dependencies and what their downstream objects are and can distribute loading tasks across a data conversion team accordingly.

Ratify programming standards, data formats and naming conventions

Standardizing source code makes that code easier to read, easier to troubleshoot and easier to recycle. Standardizing reduces the time spent documenting source code for other programmers. Standardized source code is interchangeable among programmers and facilitates the redistribution of tasks from over-burdened resources to under-utilized resources.

Standardizing formats reduces the amount of overhead required for transforming data and eliminates unnecessary manual steps. Tab-delimited Unicode data is human-readable, is compatible with a host of powerful spreadsheet and database products and is a reliable standard. Standardizing on SAP field names in source data reduces the risk of error during loading and simplifies hand-off of data to data conversion resources who are familiar with SAP. Extracting like-kind data from disparate legacy sources into a single unified data format reduces the number of steps required to transform data prior to loading and increases the level of data integrity.

Misaligned localized settings can cause catastrophic failures during a project. By standardizing the date and number formats that the team uses, many errors can be avoided and development times reduced.

Set optimistic milestones for development

Setting dates for requirements well ahead of schedule can help identify problems early and leaves time enough to resolve them before the project timeline is affected. Milestones empower the team lead to identify issues through status meetings rather than waiting for resources to complete the task with little or no guidance. Tight schedules encourage productivity, eliminate the natural tendency to take all the time available and reduce the risk of last-minute crisis situations.

Avoid manual steps

Experience shows that most catastrophic ETL failures are caused when data conversion is handled as a manual process. Steps that are not automated can too-easily be forgotten or performed inconsistently. Automating steps so that they occur in LSMW or as part of a macro in MS Access ensures that transformations are consistently applied in each phase all the way to go-live.

MS Excel is notorious for ruining data by truncating leading zeros, converting values that look like numbers into scientific notation and misinterpreting international dates and currencies. Allowing data to be massaged manually in Excel can render it useless or lead to downstream data integrity issues.

Often it seems that data is only available from legacy systems in the form of MS Excel, but in reality, Excel is used as a reporting tool for a query on an underlying database. Extracting the data directly from a database like MS SQL Server or Oracle as an automated task reduces the risk of data inconsistencies and eliminates a manual step that is often performed by somebody outside of the data conversion team.

Get it done during the day

Programmers do their best work when they are well rested. Scheduling tasks to start after business hours or on weekends is seldom necessary prior to go-live and is likely to burn out a data conversion team. Resources that are tired due to lost sleep are more likely to make mistakes and take shortcuts that can reduce data integrity and quality. Granting compensatory time off during lulls in the project is a great way to reward hard work, lost sleep and diminished social life.

Be an advocate

Simple tasks like getting an account on a server, finding a chair, procuring a computer, obtaining appropriate permissions or even scheduling a meeting room can be daunting tasks for a new resource in a large organization, especially when stretched thin on an SAP implementation. A team lead can greatly increase productivity by acting as a single contact to resources that generate, manage or obfuscate internal resources. Arranging for the fulfillment of all requirements before the data conversion resource starts working on the project allows for immediate productivity and eliminates lost hours caused by delays in account procurement and management.

Negotiating an enhanced service level agreement (SLA) with the client's internal IT department is a great way to streamline the resolution of small problems and reduce lost productivity caused by inaccessible IT resources including hardware, software and personnel. A very subtle SLA improvement is the ability to skip tier 1 support lines and call directly to tier 2 or 3. When possible, arrange for a dedicated resource in the IT department to act as the personal helpdesk for the project team. This can mean the difference between a three-minute turnaround and a three-day turnaround.

Back up conversion data and software

Too often the importance of consistent, reliable backups is understated and under-appreciated. A failed hard drive, stolen laptop or careless slip can set a project back months.

Having backups on-hand allows for immediate recovery of lost data without having to rely on the IT department. Backups can be used as a historical record of the project for the purpose of documenting progress and requirements. A hundred-dollar external hard drive might store an entire year of backups for a large SAP implementation including all revisions of legacy data extracts, transformed data, mapping documents and software-based tools.