by Emily Chan
Table 6.1 Time to Market
Step 1: writing a book (2 years) Step 2: printing (1 month) Published date
1 January 2008 to 31 December 2009 1 January 2010 to 31 January 2010 31 January 2010
1 January 2010 to 31 December 2011 1 January 2012 to 31 January 2012 31 January 2012
1 January 2012 to 31 December 2013 1 January 2014 to 31 January 2014 31 January 2014
Two technical terms are useful when trying to understand the relationship between a bottleneck and the whole process:
Cycle time. Cycle time is the time between two output units generated by a step of the process or the entire process. The cycle time of a process is the same as the cycle time of the bottleneck. The step of writing (the bottleneck step) and the entire process (writing and printing) will both have a cycle time of two years per book. Notice how the cycle time for printing is not a factor in this consideration.
Capacity. Capacity of a process is the maximum rate of output of the entire process. Capacity is measured in units of output per unit of time.
The inverse of cycle time is capacity. This means the capacity of writing and the capacity of the entire process look like this:
If process capacity is the inverse of process cycle time and process cycle time is determined by bottleneck cycle time, then process capacity is determined by bottleneck cycle time.
Once this relationship is clear, then the importance of identifying, understanding, and managing bottlenecks is self-evident.
Is the process doing what it is supposed to do?
When processes are first established, they are usually quite effective in serving their intended purpose. However, over time, processes can become ineffective if they are not updated to keep abreast of changes inside and outside the company. Reasons for failure to update could include inertia, inability to see the need for change, inability to make a change, or resistance to the changes in power bases that would result from changing the process.
One key example of ineffective process I have seen in a number of clients and HBS case studies is the outdated sales resource allocation process. This problem is vividly described in the Harvard Business Review article “The New Science of Sales Force Productivity,” where the authors used a fictitious sales manager, Bob, to summarize the issue of the old process that is used by many companies in allocating their salespeople’s time:2
Bob Brody leaned back in his chair, frowning. Corporate wanted another 8% increase in sales. . . . Ah for the good old days, when he could just announce a 10% target, spread it like peanut butter over all his territories, and then count on the sales reps (to use their personal relationships) for each region or product line to deliver. . . . Today, the purchasing departments of Bob’s customers used algorithms to choose vendors for routine buys; pure economics often trumped personal relationships. . . . Bob was overwhelmed.
Here’s a real-life though somewhat disguised example of the same problem. One of my clients, a property investment firm (call it Company P) generates revenues from rental income from its shopping malls. Company P’s share price was depressed due to below-market revenue growth. The consulting project found that the sales process was becoming ineffective:
Historically, Company P’s sales team has always been autonomous and powerful. It has been in charge of the entire process of rental from tenant identification to rental negotiation to rent collection ever since the company started.
The process was effective in the past when the market was simpler—a rental deal meant agreeing on a fixed monthly amount. However, as the market became more complex, rental deals also became more complex. For example, in addition to a fixed monthly rent, sometimes there is a variable turnover rent that is paid as a percentage of the tenant’s actual revenues if revenues for the month exceed a certain fixed minimum threshold. Or some key anchor tenants will demand an up-front subsidy or loan for renovation or a two- or even three-year rent-free period.
Company P’s retail sales team did not have the analytical capability to assess these terms. As a result, subsequent analyses showed that many of the terms they signed were unprofitable because the turnover rent threshold, up-front subsidy, or loan to the tenant was too high.
Another problem with the process was the market itself, which was getting more competitive with many more shopping malls. It became critical for Company P to differentiate by signing up unique, successful overseas retail brands and concepts. But the sales team did not have the experience of actively identifying and marketing to such retailers.
The consulting project recommended some key changes, including a process for identifying and marketing to overseas target tenants and involvement of the finance department in approval and monitoring of complex rental deals.
Is authorization appropriate?
Many processes require authorization. For example, a payment process in finance may require signatures from the department manager applying for the payment, the finance manager approving the payment, and the finance director for signing the checks. Such checks and balances are necessary for risk management. However, a process that requires authorization by people who do not have the appropriate qualifications can result in ineffectiveness or sometimes inefficiency. It is important to question and challenge each authorization on the grounds of need and effectiveness. For example, I was consulting for one of the biggest state-owned banks in China. The bank had a bad debt problem and wanted to improve its loan approval process. After some analysis, we discovered that the loan approval committee consisted of a large number of people who were very senior at the bank but who were not trained in risk management. To make it worse, many of them tried to avoid these meetings or put them low on their priority list. As a result, it was very time-consuming to schedule loan approval meetings (which of course added to the lead time for loan approval) and the decisions taken at these meetings were often suboptimal. By redefining the composition of the loan approval committee, the bank was able to not only improve the quality of the decisions but also reduce the time for loan approval (thereby improving customer service) as it became easier to schedule loan approval meetings.
PROCESS MAPPING
Process mapping is an important tool for visualizing, understanding, and improving a process. It has two key components:
Construction of process maps or flowcharts.
Analysis of key components of the process maps.
Usually two maps are drawn for each process—one before process reengineering (or process improvement), and one after. Best practice principles are applied to each step depicted in the before map to understand how the process can be improved. People sometimes refer to the analysis of key components as “off-map analysis” or “key analysis,” and I find these terms useful even though they are not official.
Types of Maps
Process maps come in three general types. The choice depends on the process involved. It is easiest to illustrate process mapping by examples.
Basic Maps
Expanding on the earlier example about my capacity for writing and publishing books, Figure 6.1 illustrates mapping that process.
Though some professional standards for flowchart shapes, such as a rectangle for a process step and an inverted triangle for inventory, have been established, I have never felt it necessary to follow any of them strictly. The key is to clearly define and be consistent in your own usage.
Process improvement can be done by applying the best practice principles to this map and starting from the most time-consuming step. These are some of the key “best practice” questions to ask:
The biggest bottleneck is text writing. Can the text writing time be reduced? Possible solutions could be hiring one more nanny to reduce distraction from my young baby, hiring an assistant to help organize research materials, or reducing the family vacation.
A major inefficiency is rework. Can the number of reworks be reduced? My publisher and I could brainstorm together on how to reduce rework, such as the publisher generating a ch
ecklist of requirements, or me submitting each chapter as it is finished so I can take the feedback into consideration as I continue to write more chapters.
Can review time by the editor be reduced?
Can waiting time at the printer be reduced?
Figure 6.1 Process Map for a Book
Multi-Party Maps
For processes where a lot of different parties (people or departments) are involved, the format illustrated in Figure 6.2 can be used.
Figure 6.2 Multi-Party Process Map
This format can highlight any inefficiency due to multiple parties’ involvement, such as excessive authorization and review, handoffs back and forth leading to waiting time and rework, and so on. This example is a simple one, with few parties involved. I have seen processes where so many parties are involved with so many handoffs that arrows were running all over the page. This usually indicates significant opportunities for improvement.
Black and White Space Maps
A black and white space map is most useful to highlight low- and non-value-added waiting time, rework time, and other delays in the process. It is also useful for identifying bottlenecks and inefficiencies. Focusing on rework review in the book publishing process produces the map shown in Figure 6.3.
A few points to note on this map:
The horizontal axis is time elapsed that affects the critical path.3 It should be drawn to scale as much as possible so it is easy to see the process visually. It has a lot of impact when the color for non-value-added steps dominates the map.
Figure 6.3 Black and White Space Map
The value-added steps are shaded in one color (black is my preference.) Some steps are obviously value-added, such as physically writing the book, approvals necessary for risk management, reports required by law, quality checks for mistakes that could not be systematically and perfectly avoided, and so on.
The non-value-added steps are shaded in a contrasting color (white is my preference.) Some steps are obviously non-value-added such as waiting time, rubber-stamp type approvals, rework for mistakes that could have been systematically avoided in the first place, and so on.
You may find it necessary to have a third category—semi-value-added. For example, if the re-review by the publishing editor after review by the senior editor is 95 percent reading the comments and only 5 percent adding further insights, then it can be seen as “semi-value-added.” Usually, I like to use gray for the steps that cannot be easily eliminated (as they are not totally non-value-added) but maybe can be reduced or streamlined to increase process efficiency.
As discussed, best practice principles are applied to the process maps to help identify areas for improvement. It is worth noting that best practice principles are applied not only after the “before” map is fully drawn, but also during the research and the drawing of the maps. For example, if there seem to be a lot of complaints about a certain step in the process, then special attention should be paid to verify (with other interviews or analysis steps) and highlight this issue in the “before” map.
Besides best practice principles, benchmarking similar processes within the company or at other companies is also a powerful tool. Benchmarking can help identify issues, stimulate creative solutions, and convince the organization of the need for change. This is especially useful if an organization is facing a new challenge and does not know how to adapt or create a new process to respond. A major liquor client of mine spent millions of dollars a few years ago to benchmark strategy-planning processes of leading companies around the world. The key reason was the CEO’s decision that the company’s top-down strategic planning process, whereby each business unit was given a target to be met, was no longer sufficient as the differences and the level of competition of the different markets increased. He needed to understand how leading companies did their strategy planning in order to design a new process and to convince his organization of the validity and need for the change.
An analogy for process reengineering is plastic surgery. There are usually two maps: before and after the surgery. To decide what to do in the surgery, classic universal beauty standards are available to assess the “before face”: wrinkle-free skin, straight nose, and the like. But sometimes, perhaps because the patient cannot describe what the new face should look like, then benchmarks like “the lips of movie star Angelina Jolie” will be used. Finally, some analysis may need to be done in addition to the exterior of the face, for example, bone or muscle structure of the face, or levels of blood pressure, as part of the surgery. Such analysis will be analogous to the “off-map analysis” to be discussed in the next section.
“Off-Map” Analysis
Process mapping is an important tool for visualizing and defining the steps in the process. However, often, more in-depth analysis needs to be done on certain steps of the process. The analysis focuses on understanding and resolving any issue found in the process map. The analysis often results in recommendations for tools, systems, or automation.
Simple Analysis
Figure 6.4 shows an example of simple analysis: understanding the relative severity of different causes of inefficiency. Using the book example in the basic map section, the bottleneck is the step “writing the text.” To determine how to reduce the time needed, off-map analysis can be done to dive into this step of the process.
More in-depth analysis can be done on each of the items, such as what is taking up so much research time. Solutions such as outsourcing or automation can then be identified to reduce the time.
Figure 6.4 Bar Chart for Simple Analysis
More Advanced Analysis
Sometimes, when you have enough data, more advanced analysis can be done. A very good example was given in the Harvard Business Review article “The New Science of Sales Force Productivity.”4 Michael Pilot, then president of a unit of General Electric, was trying to solve the sort of client prioritization process problem highlighted by the fictitious Bob mentioned earlier on:
Problem Pilot faced: “The company’s field sales managers even manually classified all the names in the division’s database as either high priority or low priority . . . (relying) on telephone books . . . newspapers . . . signs on trucks as they went by or signs on buildings . . . (Pilot) knew that GE Commercial Finance had to ‘put some science into it.’ ”
To improve the account prioritization process: “Pilot asked his field managers to create a list of . . . criteria that they believed would correlate with the customer’s likelihood of doing business with GE. He took the 14 features they came up with, ran regression equations against the database of transactions and identified six criteria that had high correlations. If a prospective customer tested well on those six criteria . . . the probability that it would do business with GE is high.”
Improvement: Pilot said, “We found that the top 30% of prospective customers were three times more likely to do a deal with us than the bottom 70% . . . yet only about half of them were previously classified as high priority by sales managers.”
Data Sources
The data sources for process maps and analysis are very similar to the data sources for strategy study discussed in Part III of this book. Here, it is enough to highlight some of the key points:
Interviews are a key source of data. Very few companies document their processes in detail. Even when they do, what is on paper may sometimes be very different from actual practice. Hence, interviews with people involved in the process are a key source of data, especially for the maps.
Quantification is important. Quantification measures the extent of any inefficiency or ineffectiveness, helping you see how big any problem is and hence the priorities, and convince others of the need for change. For example, if rework is high, you would want to know how many times things go back for rework, how much time each rework takes, how much material is wasted, how much money can be saved if the process is changed, and the like.
Sampling is a useful tool. Interviewees sometimes give you conflicting descriptions of
the process, or find it impossible to give you the information you need. This can be because they are trying to hide issues or because the process is truly messy. In such cases, one tool I have found useful is to do sampling, as described in more detail in Chapter 12.
Identifying root causes of any issue is important. This is self-explanatory. The key skill is “asking why five times,” and is discussed in Appendix B.
Notes
1. Ann E. Grey and Leonard James, “Process Fundamentals,” Harvard Business Review (September 2007).
2. Dianne Ledingham, Mark Kovac, and Heidi Locke Simon, “The New Science of Sales Force Productivity,” Harvard Business Review (September 2006); quote on p. 1 of article.
3. Critical path is the process or the sequence of steps that determine the process time. For example, in the book writing example, there may be other activities like payments and meetings. But these activities are not on the “critical path” as they do not affect the process time.
4. Dianne Ledingham, Mark Kovac, and Heidi Locke Simon, “The New Science of Sales Force Productivity,” Harvard Business Review (September 2006); quotes in this section are from p. 2 of the article.
7
HUMAN RESOURCES
AIM TO BE THE STUPIDEST BOSS
If you ask the HBS graduates who are business owners or hired chief executives (and in fact any seasoned business owners) what the single most difficult thing in managing a business is, the answer would most likely be “people.” Hiring, training, motivating, and retaining people are extremely important but challenging. To me, the best business is one that generates income without any employees. That’s why I like writing books and investing in real estate and stock markets. These lines let me work largely on my own.