Sustaining Low IT Cost

Today, many CxOs believe IT cost can be managed primarily by reducing the rate of labor.  However, this is often not the case because doing so does not give due consideration to measures of productivity - that is, considering cost in relation to output. In this context, productivity becomes a significant contributor to the overall cost function. For example, reducing labor cost by 25% while reducing productivity by 50% actually increases cost by 12.5%.

To understand this better, let's consider cost to be the product of three components:

• The amount labor (usually in units of time).  There are two types of labor: Development and Management
• The rate of labor (usually currency over time, such as $100/hr)
• The productivity: for the time being, a dimensionless value through which we can express the relationship between work and output.

In this framework, cost can be expressed mathematically as follow:

Cost = (Rate x Labor) / Productivity

Since productivity is commonly associated with the labor metric (duration of work), we can further refine the mathematical representation to an equivalent one that reflect this convention:

Cost = Rate x (Labor / Productivity)

This representation allows us to clearly see that as productivity goes up, cost goes down even as rate and labor remain fixed.

Two IT Cost Reduction Strategies

Cost, as a function of labor cost and labor amount, can therefore be manipulated in basically two ways: Changing the rate of labor or changing the amount of labor (i.e. improving productivity).  This admits two primary strategies for cost reduction:

1. Reduced Rate (Labor Arbitrage)
2. Reduced Labor (Automation)

Labor arbitrage is premised on the availability of lower-cost labor available in developing markets.  Expensive domestic labor is simply displaced by inexpensive foreign labor.  This strategy is attractive, for example, if requirements are clear and complete, leaving solution design and construction to be performed by lower cost labor. Alternatively, solution design may be performed collaboratively using both in-house architects and outsourced architects. The resultant technical specification is then passed to laborers whose coding tasks are now considered relatively straight-forward and therefore require only diligent management to ensure schedules and costs are controlled.

Following conventional wisdom that such displacement of labor reduces cost, executives set targets for the percentage of IT work to be performed offshore. One notable consequence of this strategy is that, over time, much of the knowledge work fades from the organization. The design and construction aspects of the solution development cycle are reduced to simple vendor management with little ability to innovate or add value. The final outcome is a pure "back office" operation outsourced as a non-differentiator much like Finance and Human Resource in many organizations today.

Reduced Labor or Automation eliminates the need for menial and repetitive labor that tends to characterize the solution construction and hand-coding portions of the solution development cycle. The knowledge-based, high value work is retained by the organization preserving its ability to innovate and add value.  The IT organization becomes part of the "front office" value chain driving and enabling revenue instead of a "back office" cost center managing vendors.

This thinking heralds a new approach to solution development where knowledge is captured in rich requirements models and logical design models through a collaborative partnership of business and IT knowledge workers. These models are then automatically mapped into complete solutions that execute on any platform in any environment. The technology required to map models into complete solutions is provided by a new generation of outsourcing partner whose business model is based on providing both models and low-cost software components that can be declaratively incorporated into solutions.

These outsourcing partners enable businesses to focus on market-separating business concerns while providing models for commodity business components. In some cases these partners can also provide technical assets (e.g. transformations, mappings, transformation engines,etc.) that convert models into executable solutions. The key differentiator for these harbingers of a new outsourcing model is the focus on Reduced Labor (Automation).

The figure below illustrates the contrast between labor arbitrage and automation.  Consider a project that is estimated at 10,000 hours.  If the cost of labor is $80/hr, then the project cost would be $800K.  If cheaper labor ($60/hr) could be used to perform 80% of the work (first row), the cost could be reduced by 18% (a savings of $150K).  If, on the other hand, automation could be used to eliminate 50% of the work, then the cost could be reduced by 50% (a saving of $400K).  If automation could be used in conjunction with cheaper labor (second row), then the cost could be reduced by 68% (a savings of $550K), even though more hours of the project would be performed by the more expensive labor.

Managing Labor Cost

Not only is automation superior to labor arbitrage as a cost management strategy, it is far more sustainable because it leverages higher productivity over cheaper labor.  On the other hand, organizational productivity diminishes within the labor arbitration approach.

Sustainable Cost Management

There are two things to consider regarding labor arbitrage as a strategy for cost management:

1. The cost of offshore labor increases more rapidly than the cost of domestic labor
2. The software development effort increases due to management overhead and technical oversight

Offshore labor rates often increase exponentially compared to domestic rates because the influx of wealth to any economy inflates that economy.  As inflation rises, so does the cost of living, which in turn increases the cost of labor. This means that labor arbitrage as a cost management strategy cannot be sustained.

Additional hand-offs, communication issues, and frequent "black box" approaches to labor arbitraged outsourcing also contribute to the Project overhead. Beyond that, gains are further offset by the simple fact that there are now two organizational infrastructures involved, including more managers and more reviewers. While there are developing markets that have excellent educational infrastructures, it bears mentioning that the software development field has one of the lowest barriers to entry of any professional field.   that Since cheaper labor is often accompanied by lower skilled workers, there is a need to have more managers overseeing more junior level workers. There is also more work involved to verify the quality of the product (quality assurance) and to correct for it (rework).

To explore the implications of this, consider the two graphs below depicting a Traditional IT model, a Reduced Rate IT model, and a Reduced Labor IT model.  The following assumptions have been made to illustrate the point:

1. It is possible to cut the rate of labor by 50% with labor arbitrage (25% is the norm today).
2. Offshore development adds 10% labor overhead (more management and quality assurance).
3. It is possible to cut the amount of labor by 50% with automation (75% is the norm today).
4. Labor reduction using automation technology adds no labor overhead.

These curves point out the following:

• Labor reduction (i.e., automation) reduces cost more than rate reduction (i.e., labor arbitrage).
• The gap in cost reduction closes over time for rate reduction relative to the traditional IT model.
• The gap in cost reduction increases over time for labor reduction relative to the traditional IT model.
• At some point, the offshore model becomes more expensive that the traditional model because of increasing labor rates in offshore markets and increased project overhead.

Conclusion

Clearly, the offshore model is not sustainable.  It requires a continual quest for lower cost labor markets as today's lower cost labor markets continue to diminish due to labor arbitrage.

On the other hand, the automation model is infinitely sustainable, insulated from rising labor rates, and dramatically accelerates the rate at which the critical contributions of the knowledge worker can be realized as solutions.