Online retailer ABC sells books, DVDs, CDs, MP3 downloads, software, video games, apparel, furniture, food, toys and jewelry. ABC has a strong market research unit and a seamless feedback loop for enhancing the customer experience. In the last few years, however, sales had dropped significantly contributing to a slip in the overall market share of ABC.
The CEO of ABC asked his executives, If we are constantly collecting market and customer feedback, why are our sales so low? Especially given that we have highly productive teams and that our market research team was recognized as the best in the industry last year.
John Smith, an independent Lean Six Sigma (LSS) consultant was invited by the CEO to guide ABC in improving their declining sales. Smith spent some time talking to key personnel from different teams; a few of the teams shared statistics around the productivity of their individual teams as well.
The consultants first observation was that each of the teams within the company operated in a silo. Each team reported their capacity, productivity and output individually. This meant that the end-to-end process of translating customer feedback to a better online customer experience was not being seen as an enterprise-level process. For example, e-Strategy responded to the findings from Market Research in isolation of the Product and Service teams.
One of the companys executives asked Smith, Is isolation really an issue? What does it have to do with low sales?
Smith explained that when the teams respond independently, they dont have an opportunity to review each others data and weigh in before a recommendation is sent directly to Solution and Delivery to get implemented. It is possible that a solution devised by one team could impact the work of another team. Without coordination among teams, their isolated efforts could lead to rework, overproduction or other wasted work thereby contributing to a decrease in the percentage of customer feedback items being translated into features.
Prioritizing customer feedback with respect to multiple parameters such as value, risk, cost of implementation and speed to market is only possible when there is a single, organizational framework to manage improvement activity. Smith proposed that the productivity of the teams should be measured based not only on output but also on outcomes.
Consider two of ABCs website issues: 1) inconsistent branding and 2) performance issues due to an increased number of concurrent users (leading to people dropping out before finishing a transaction). Imagine also that it is four weeks before the Christmas shopping season begins a time when the number of concurrent users increases by at least 50 percent.
The Marketing teams priority is to get the branding correct and may be able to get those requirements to Solution and Delivery quickly without talking to Operations. But why worry about online branding if the website is already unable to cope with the existing customer traffic? Parallel processing generally focuses on individual team outputs; with an overarching improvement framework, priorities become clearer and the focus can be shifted to customer outcomes.
Smith described a hypothetical situation in which LSS governed at the enterprise level the overall process of translating customer feedback to a better customer experience. The big business Y in this enterprise-level process is decline in sales figures.
There are likely multiple causes contributing to the Y, Smith noted. Without having performed detailed analysis, he suggested looking at productivity and output. Various ABC teams had spoken to Smith about their individual teams productivity as a positive metric; they were delivering a significant amount of output. The incongruity between the perceived high productivity and the declining sales led Smith to want to investigate further.
Smith shared Figure 1 with the executive team. The process shown is that of the information technology (IT) division, which helps translate feedback from online customers into product features. There are multiple stakeholders involved in this process, ranging from the survey company (external to ABC) to the internal marketing team, and from operations to the IT teams. Smith noted, This process cuts across 10 separate teams. I have assumed productivity of about 90 percent for each team. This in Six Sigma terminology is the unit yield for a team. That means, irrespective of the total capacity of the team, 90 percent of the customer feedback that a team receives is incorporated into website enhancements or some process change.
Smith directed his next comments to the CEO: This is what you see in all performance dashboards. The usual misconception is that if all the teams have a productivity of 90 percent, the average productivity for the entire process should be about 90 percent. In effect, what that should mean is that 90 percent of the feedback captured during market research translates tangibly into improvements that reach your customers. So if that was the case, why are the sales so low? The reality, however, is that only 35 percent of the feedback captured by market research translates into some tangible action. And in most cases the feedback is outdated by the time any action has been taken. Look at what happens with a mathematical model when all of the contributing teams have the same productivity.
Productivityenterprise level = (Productivityteam)number of teams or
Yieldoverall = (Yieldstep)number of steps
Smith continued to address the executives of ABC: “In this example of a sequential process, the productivity (unit yield) of individual teams is about 90 percent, but you have to multiply them to get an accurate measure of enterprise-level productivity (final yield). They cannot be averaged. Think of the first two teams in this process. The process begins with 100 customer feedback items at the Market Research team and 90 pass. The starting number for the e-Strategy team is then 90, not 100, and 90 percent of them pass, leaving 81. The average team yield is 90 percent but the final team yield is only 81 percent. This is how the full IT process has a predicted final yield of only 35 percent, resulting in customer dissatisfaction and poor sales.
This sequential example highlights the effect of measuring productivity at the team level without having an enterprise-wide view. Of course, the problems that ABC faced were multi-dimensional there was more than one cause contributing to the Y of declining sales.
Productivity often does not fully reflect the cost dimension of a problem, Smith explained to the ABC executives. Operating costs also affect a companys bottom line even if sales increase. By not having a cross-functional team or enterprise-wide process management, the unnecessarily large number of steps and teams can lead to higher operating costs related to failure, overproduction, rework, etc.
Although increasing sales is pivotal, optimizing operating costs is paramount in order to assure sustained profits. To understand the costs of ABC, it is important to examine the journey of the items leading to the 35 percent final yield.
Consider two teams in an overall process through which 100 items begin with Market Research and then need to pass through Product and e-Strategy. Assume the cost of operations of the Product team is $X per year and $3X per year for the Strategy team. In addition, assume that each team has 10 critical-to-quality characteristics to meet. The individual team yield is 90 percent for both teams so there is the possibility that each team will produce 10 defects in 100 defect opportunities. In terms of their ability to produce defect-free quality characteristics, they are at the same level.
But what is the cost of fixing such defects and the costs of rework or additional handoffs? If the costs are $Z for the Product team, then they are $3Z for the e-Strategy team. In this case, improving the productivity (throughput yield) will help improve the sales numbers, but it is critical to look at optimizing the operations to maintain long-term viability. Having a high number of teams with varied costs of operations combined with the lack of an overarching improvement framework put the company in a losing situation. This situation can be mitigated by having cross-functional teams work on a prioritized set of items delivering incremental value.
Smith suggested that ABC enact an enterprise-wide transformation program to improve sales and profits using LSS. The key areas of focus of this transformation would include team rationalization, review of operating cost bases, a review of the current delivery methodology and reporting. In each of these areas, Smith applied the LSS filters to identify and eliminate waste as well as improve the quality at each stage all contributing to an increase in the overall throughput.
In the end-to-end value chain of ABC, there were 10 different teams. When looked at with a Lean filter, it was obvious that a significant amount of handoffs, rework and motion contributed to a low throughput. Smith recommended forming cross-functional teams by identifying the groups with similar objectives. Figure 2 shows the value chain before and after this rationalization exercise.
The rationalization resulted in a 40-percent reduction in the number of individual teams. This was a significant step for ABC as it reduced the number of non-value-added steps in the overall process and facilitated greater collaboration among team members. The reduction in the number of teams resulted in fewer handoffs, leading to a seamless flow of information.
The existing delivery approach in the IT department of ABC was a traditional waterfall methodology, a sequential design process used in software development in which progress is seen as flowing steadily downward (like a waterfall) through the phases of conception, initiation, analysis, design, construction, testing, production/implementation and maintenance. In the initial scenario with up to 10 teams, following a waterfall delivery created an issue with the speed to market. Apart from speed to market, most of the supposed value propositions seem to become outdated by the time they were delivered.
With fewer steps and teams in the value chain, the speed to market will improve, but further improvement is still needed. In the ideal state, ABC would deliver against a prioritized set of feedback items (prioritized against such considerations as value, risk, cost of implementation and speed to market) that are not already outdated, using a seamless feedback loop.
A key limitation with the waterfall process is the expectation that all of the best ideas have to occur at the beginning of a project. In reality, good ideas can appear at any time. A rigid, change-resistant process (even if leaner) tends to produce mediocre products, sometimes at a point in time after the customer has moved on to another company. Achieving maximum throughput, even in a Lean organization, requires each step to deliver as much value as possible at regular intervals all along the process.
By adopting an agile delivery approach, which is more flexible than the waterfall approach and follows an iterative process to deliver value quickly, the productivity of each ABC team was further increased. Using agile meant that the cross-functional teams worked collaboratively to prioritize the items that were identified by the market research team and started to deliver upon those priorities in chunks. The teams delivered more value incrementally and were able to respond to market changes dynamically. The other key benefit of this approach was that an item identified in the most recent customer survey could jump onto the priority list if it was deemed to have the potential to deliver greater value than other priority items. Introducing new changes were easier and the products became more relevant to the customers of ABC at a faster pace.
The recommendations for ABC can be summed up in LSS terminology by the figure in Figure 3. The y-axis is the number of teams that were part of the initial value chain. The rationalization of the teams (the reduction from 10 to six) was achieved by eliminating waste using Lean. The adoption of an agile approach (as opposed to the existing waterfall approach) improved the quality and productivity of each of those steps and teams. Lean and Six Sigma complement each other and tackle problems multi-dimensionally. Lean helps reduce non-value-added steps in a process by eliminating rework, handoffs, motion, etc. Six Sigma works relentlessly to improve the quality of the value-added steps, which contributes to reliability at the enterprise level.
The other key recommendation to ABC was the acknowledgement that each team has a different cost of operation and, therefore, a different associated cost of failure or rework. By creating cross-functional teams with representatives from each of the departments, the total cost of failure was also reduced.
LSS provided a structured approach to eliminate waste, reduce the cost of failure, and improve the throughput yield and speed to market of the overall processes for ABC, resulting in higher end-customer satisfaction. LSS also helped to deliver products in a manner that optimizes resources and increases value to meet customer expectations. This is exactly how Lean and Six Sigma in combination can drive greater benefits at the enterprise level than either alone.
For ABC, the number of handoffs and rework were significantly reduced and contributed to an increase in individual team yields. The pace and relevance of value delivered by ABC also increased as the teams worked against a prioritized set of feedback items, focusing on producing incremental value. Customer satisfaction increased and so did the sales.