Controlling IT Costs; Enterprise Architecture (EA) strategy, a shared lexicon, and enforced change
To control Information Technology (IT) costs we think about and act within the enterprise as a whole, in part because we sell enterprise and mid-level solutions. We apply an Enterprise Architecture (EA) strategy which at the top level is comprised of infrastructure and communication considerations. This is not just about technical infrastructure, defined or designed by IT, because it is highly likely that such individual solutions (one offs) will not align to core business strategies (vertical needs verses horizontal needs spanning the whole company).
It is not really possible to do this, that is consider the entire company's needs, without significant participation by the business for which we use terms such as Solution Delivery or Product Management. Product and program managers from a solution delivery framework gather information, report back to the business, and return to apply the business strategies to align with short, medium, and especially long term business goals.
This business and implementation strategy focus is a change agent, to reduce siloed thinking, and achieve more horizontal capability across units. We reduce multiple applications, which take time to manage and maintain, and where it makes sense, fold them into one. Because we take security and privacy of our customers very seriously, any applications which may be at risk have been identified and are brought up into our standards. The process of combining risk management goals, application and data reduction streams saves money, although the process of so much change at once can be stressful at the unit, project, and personal levels.
We seek to empower self-service among our partners, customers and employees, for access to all kinds of information they need, and internally reduce redundant data stores, for example referring to customers by one identifier if possible. This is especially challenging in our partner relationships with multiple data stores that contain similar information about customers which are identified in completely different ways. This is the reason for serious data modeling and tight or loose coupling where needed – to retrieve and move information back to the partner systems. We leverage Microsoft software, and then buy, build, minimize or reuse existing systems.
In order to be more successful in our efforts to control IT costs we strive to increase flexibility among existing staff and provide rewards for strategic thinking – this strategic thinking aligns along company-wide goals. We need people with the right skills who work in efficient methods, only including the people who need to be included to make decisions or act. In fact we need to change confrontational and passive aggressive behaviors internally to collaborative personality styles – changing the organizations culture is doable but difficult. For more information I recommend reading "The Heart of Change" by Kotter and Cohen.
The technologies we invest in to help control IT costs are our own. We custom write stuff served up on Microsoft servers and plan to use SharePoint as the UI for our new change request tool. We are substantially reducing and eliminating the number of different applications (SQL stored procedures or XML Blobs mostly) we use and maintain on a daily basis. We are moving from C++ to C#/.NET (C Sharp and .Net technologies).
We use Microsoft software as our strategy to control IT costs - it is easy to manage, and has great support. Some team members keep an eye on relevant Open Source software as competitive analysis.
Our company is getting the maximum value from its data center investment because we have not invested to the level we need for our infrastructure. We expect to remediate this lack of investment after deploying skilled, thoughtful product managers with the right combination of education and practical experience to assist in this effort through the next couple of years.
What is our organization doing to maximize the value from its data center investment? In addition to the other things mentioned we outsource development and support to India, Israel, and developing countries, etc. We also are making use of tax advantaged locations for large savings in transactions.
We are adding metrics and measurements by which we evaluate not just personal progress but internal and external customer satisfaction with our IT initiatives on a project by project basis to self-improve.
The practices which enable us to maximize value from our IT investment are varied and multifaceted. To maximize ongoing investment we are adding solution delivery strategies, planning ahead, and aligning IT with company-wide goals. Of course in our space we have some unique issues, and as a public company even more so. One thing that may surprise you is some of our projects we do end to end locally because of how critical success is. We leverage our best, most successful local managers to produce projects and design larger scale solutions if we determine it is the best strategy – so in this way we are flexible – we don't just out source everything.
We are in the process of reducing the number of applications we need to maintain, and where it is appropriate fold one into another so long as the user interface or back ends do not become unmanageable. We are making over our change request platform from top to bottom which we feel will enable quicker turnarounds on change requests – it is both loosely and tightly coupled where it needs to be. For the presentation layer we choose Microsoft SharePoint.
Conversely, what factors are inhibiting our organization from reaping the maximum value from its data center investments? The factors inhibiting the maximum value include a lack of foresight in strategic planning for long term goals –
1. Putting temporary things together to just meet immediate needs.
2. Focusing on small details and not seeing the big picture.
3. Lack of metrics to evaluate progress, process, and client / customer / partner success.
4. Unwillingness of team members to change or promote change even when it is in their and the companies' best interest.
5. Having too many data centers, identifying customers in too many ways.
How important is productivity within the IT function in our efforts to control IT costs and maximize our data center investment? Functionality, capacity, and reliability far outstrip productivity, but that is only because we have already hit very high productivity goals and exceeded them. Here are some of the metrics we examine:
Percentage of project budgeted costs
Total cost of ownership
Defects rate (sev1, sev2, sev3 bugs - zero tolerance for sev1)
Customer satisfaction scores (cust sats)
Flexibility of management styles
End-to-end throughput time per client-side user request
Defects per thousand lines of code (KLOC or by function)
Support functionality and documentation availability, and completeness prior to launch
Rates of failure
System acceptance (signoff)
Average turn around time for service and change requests
Number of security or privacy defects (last two should be zero tolerance in launch candidates)
Number of post freeze change requests
Among the mandatory metrics used are peer review effectiveness of code, and post mortems and overall customer satisfaction. In other words we do not consider just ontime delivery of products, enhancements, or new functionality.
What is our organization doing to improve productivity within its IT function?
Getting the right people – some people grew with us or came to us with deep knowledge from the school of hard knocks – work experience – we seek to capture the most knowledgeable and either increase their education or find those with both practical work experience and advanced degrees. Good thing this is Seattle with its heavily educated population. New programs at the university level such as Informatics and Information Management are producing the people we need – not just MBAs or Master of Comp Sci - because so much of our development work we outsource to India and developing countries, and IT is not traditionally closely aligned with marketing or sales. We do outsource much of the development work as is possible.
The undergrad Informatics and Master of Science in Information Management programs at the University of Washington are housed in Mary Gates' Hall, renovated and named in honor of Bill Gate's late mother, it's headed by Mike Crandall (Dublin Core, Microsoft, Boeing). So you can see this is the direction we are going regionally, because that is where the spend is. Another great information school is at the University of California at Berkeley, housed in one of the oldest and most architecturally beautiful collegiate buildings on the west coast, South Hall. On the physical level all Berkeley had to do is add wireless. Excellent academics such as the seminal thinker Dr Michael Buckland are there at Berkeley, and business leaders such as Mitch Kapor. Industry wide I think iSchools are having an effect, adding a more well rounded, even playful culture to high tech operations.
Improving and opening the culture is important. Having a shared lexicon is one of the benefits of educated people; those with MSIM (master of science in information management), Informatics, technical MBA degrees can comunicate effectively with highly technical people - this can produce enormous savings and long term cost benefits. Increased, clear, enthusiastic communication saves IT costs.
In strategy meetings, for example, we often include Enterprise Architects to assist in stack ranking program and project development, because this helps reduce redundant systems.
Our organization's ability to measure the return on investment (ROI) or success of its IT investments is “Fair but mixed,” we want ROI to be easily measureable and this means evaluating the correct things, asking the right questions in the first place, not following other organizations techniques, although we examine them as examples.
We are adding ways to evaluate our ROI – we do use business analysis methods. There is always an identifiable way to analyze and measure the relationship of what something costs even if it appears intangible such as Brand protection.
Considering the strategic and tactical stuff we are doing, at the core, creativity is what drives our success. Creativity is always a very difficult thing to measure. In fact it could be said that if you try, you are barking up the wrong tree. However creative thinking around practical goals has provided us success. This is where the ideas around flexibility and being very responsive come to play.
We have found very very high ROI around outsourced projects because they must be clearly defined within the Software Development Lifecycle (SDLC) and Sarbanes-Oxley Act of 2002 (SOX) compliance.
Those people who actually think out of the box are oftentimes not recognized by co-workers and management. Change is perceived as negative among full time staff. We seek to show support for both full time employees and consultants, and change this view and enhance their ability to communicate ideas. That is why our management keeps an open door policy. Unfortunately like any other policies the hazard is that individual managers must believe in our policies around openness and creativity; such self-selecting polices are impossible to enforce.
Our organization uses balanced scorecards, Six Sigma and other types of internally derived quantitative value measurement methods to measure the ROI or success of our IT investments.
The continued use of these methods we expect will substantially improve the management and measurement of our IT investments. Some of the metrics are at the discretion of the product or program manager, others are mandatory. In part we have some success- at issue is adopting metrics and measurement as well as Enterprise Architecture and engaging with open arms increased strategic thinking and planning.
Senior management must come together and present a unified strategy for the entire company – which is a top down management style but it must be embraced from the bottom up. This is within a framework of enforced change as we seek to achieve excellence in all of our business units, especially in core infrastructure – those units which either produce money, or cost money. Some of our key investments we know are lost leaders, but other research will more than make up for those. Enforced change in this context means business units receive minimum budget until they comply.
We are still feeling the effects of the changes the Web brings in enterprise directly and for our customers; we continue to learn from the effects of communities and communication via the Web. The opportunities for growth are so enormous that it is all the more important that we curb spending where it is not required and apply it as much as possible to grow in creative arenas which still have huge untapped profit potential. It is not just about money, among hard core technologists – those who really love it – money is secondary in many ways - it’s about the fun stuff technology can bring as well as the benefit to serve humanity that technology brings.
High tech, information technology, and software development have made some strides to maturity but we are still learning new things; it will be a learning industry, discovering and inventing stuff for a long time to come.
Enforced Change is a radically different challenge, and promises different ways of looking at human-to-human, individual-to-corporation, corporate-to-corporate, human-to- computer interactions, etc, which I plan to cover in future articles, so stay tuned!