Who’s Fault Is It If Your Integration Factory Is Not Performing?

The CIO of GT Inc. (the fictitious name of a real company) met with his middleware vendor rep to deliver some depressing news.

“We established an outsourced factory delivery model two years ago using the productivity tools that you sold us and we made it our enterprise standard.  The factory results however, are discouraging use of your integration platform. Projects are not getting approved by the business because of high costs, or else project teams are working around the standard and building hand-coded solutions. Did I make a mistake in buying your software?”

This blog article is a real-life case study of a Lean Value Stream Analysis that was conducted by Informatica Professional Services (IPS) at GT as a result of the CIO’s concerns.

GT is a telecommunications company with several IT factories (outsourced consulting service providers), two of which provide integration solutions. Costs were increasing despite GT’s investment in the latest productivity tools and a strong contractual agreement with the factory vendors with strict service levels and extremely low billing rates.  A team of consultants was assigned to assess the situation to find out why the expected productivity improvements and cost savings were not materializing.

The assessment project was challenging for a number of reasons:

  1. GT is very siloed with no clear ownership for the end-to-end process; all teams considered themselves “green” in terms of meeting their SLAs yet the business owners view IT project delivery as “red”.
  2. There were no useful metrics to compare projects or teams; how could the team gather metrics that may embarrass individual teams or managers.
  3. Two different consulting firms (one for data warehouse factory and one for the application integration factory) use the same middleware platform, however their design approach differs which makes it difficult to do comparative benchmarking.
  4. GT’s assumption is that the high cost of integration is the fault of the factory vendors; but what is the real cause and how best to tell GT management that they are wrong?

To address the first two challenges, the IPS team used Value Stream Mapping (VSM) to model the flow of materials and the flow of information in an IT project. The VSM was used to calculate baseline metrics including a Value Add Ratio and Lead Time.  It soon became clear that GT managers and project staff didn’t really understand the end-to-end process – only their individual process step. The VSM also modeled the governance, approval, review and status activities which helped to shine a light on causes of delays and high costs. The GT staff that participated in the VSM exercise identified more than 10 cross-functional process improvement opportunities.

To address challenge number three, the IPS team used the Integration Point Analysis technique which I wrote about in a recent blog. By standardizing the definition of deliverables from the two different integration factories, it became possible to benchmark and compare the results of the two factories.  Interestingly, both factories performed quite poorly – if there had been a big difference in performance between them it would have been easy to point blame, but something else more systemic that affected both factories was the root cause.

The real root cause problem was an inefficient (broken) end-to-end process. A typical integration project, even very simple ones, involved 10 internal GT teams, all operating as independent silos. Every GT team was measured only on their deliverables and discrete SLAs. For example, the integration solutions team had a 5-day SLA to prepare the RFP document and the factory then had a 3-day SLA (for a simple project) to respond with a proposal. The metrics encouraged everyone to meet the SLA without regard to the quality of the RFP (which often had only very high-level requirements that were insufficient for an accurate estimate) or the quality of the Proposal (which often just used high-level rough estimates since the project requirements were so vague). Over time, team members, who of course knew that the process was broken, had stopped raising concerns since no-one had accountability for the end-to-end process so if anyone raised an issue about it,  they were perceived as complainers and told to “just do your job.”

The Lean VSM analysis was a success in that the findings and recommendations were accepted by senior IT management, Factory (supplier) management, first-line GT management, project analysts, developers, architects, Project Managers and the CIO – without assigning blame. The Lean Value Stream Maps and fact-based analysis refocused everyone on the ultimate customer (business sponsor) of integration projects, and on optimizing the end-to-end process rather than on individual silos. The plan recommendations were accepted and staff were assigned to begin immediate implementation of the identified improvement opportunities.

In short, the value stream analysis was an amazing success!  To learn more about these techniques and apply them to challenges in your enterprise, visit https://community.informatica.com/solutions/1618 to learn how you too can achieve Lean and Agile Data Integration with a Value Stream Workshop that is customized to your organization. Or simply contact Informatica Professional Services.

This entry was posted in Business Impact / Benefits, CIO, Data Integration, Integration Competency Centers, Professional Services and tagged , , , , , . Bookmark the permalink.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>