The State Of Workforce Management… And Those Responsible For It

It was pointed out to me recently I’ve written 100 Linked In posts. That calls for an assessment of whether anything of value has been offered or if it just involved filling the digital space.

At least no trees had to be sacrificed.

I have decided to create this post by using some of the themes covered over the last two years, but focusing on what those responsible for workforce management should be competent to do well and what they should actually be doing.

  • First, they should possess business acumen. 👔 In order to serve an organization it is necessary to address the issues it faces and to contribute to mission fulfillment. If one does not understand the business of the organization this is not possible. Practitioners in the Human Resource function must know what the organization needs and be able to help meet those needs. But so must every manager, supervisor and lead person, and each must recognize their responsibility for workforce management and to invest the needed effort on making their people effective.
  • Second, they must base recommendations/decisions on an analysis of all relevant evidence. 📊Historically this meant applying the skills and knowledge gained through education and experience, supplemented by an appropriate dose of intuition and creativity. Of late the emphasis is on the use of workforce analytics. Certainly the use of artificial intelligence and machine learning can augment human judgment… but it cannot replace it. Applying analytical tools to databases can disclose relationships, such as correlations. But these correlations may be statistical artifacts… sufficiently tortured data will confess to anything. And the need to innovate in today’s competitive world may limit what analytics can do. Trying to predict the future based on data that is limited to what was and what is a high risk endeavor. The effective integration of technology and people has been a theme since socio-technical systems thinking came on the scene decades ago… and it has become even more critical with the advent of new technology.
  • Third, they must understand cognitive bias and its impact on the perceptions of employees. 👨‍💼 Research in Behavioral Economics and Neuroscience have disclosed almost 100 types of bias that are within us all. Though bias removal cannot be done surgically managers can recognize what impact biases are likely to have and can create a dialogue with employees that may lessen the unwarranted employee dissatisfaction biases cause. For example, we all think we are better than we are. This impacts employee perceptions about what their contributions are and what their rewards should be… and their perceptions are their reality.
  • Fourth, they must understand the impact of cultural diversity in workforce management. 🙏Even organizations operating in one U.S. location will be likely to have employees who originated somewhere else in the world or be the product of another cultural heritage. Cultural diversity will result in different values and beliefs, which will impact what employees believe to be appropriate. The widespread focus on individual performance in the West may be seen as wrongheaded by people from a more collectivist culture. Hierarchical, top-down management may be viewed as inappropriate by those who are more egalitarian in their beliefs. Managers must recognize cultural differences when they exist, they must respect that people have a right to hold different beliefs, and they must reconcile the issues caused by cultural differences.
  • And fifth (not going to do ten), they must understand: 1. The scientific method, 2. What makes research sound and relevant, and 3. Statistical and mathematical tools relevant to workforce analytics. ⚙️This may require going back to school, but there are a wealth of online training options that can be used to augment knowledge and skills. An example of how critical this knowledge is would be the recent proliferation in the pop literature of claims that extrinsic rewards destroy intrinsic rewards. These claims are based on lab studies, such as a popular one involving throwing tennis balls at targets. Even though a lab research study is well designed and passes the internal validity test it may not be generalizable to the field (have external validity). That means the context within which the study was done must be consistent with the context within which one tries to apply the research results to. This attempt to generalize lab findings to the field is a failure and results in wrong conclusions…there is extensive field research that contradicts the lab result. Without the necessary knowledge to evaluate the validity of research practitioners may draw erroneous conclusions and make poor decisions about workforce management. The fact that so many read material espousing flawed research is evidence that there is inadequate competence in this area.

One additional thought on the debates that arise relative to sound workforce management. Most of them should not be debates at all. We live in a Quantum Physics/Fuzzy Logic kind of world today, where “both – and” replaces “either –or.” The current proliferation of articles debating whether year-end performance appraisals or continuous measurement and feedback are necessary could all be combined into one that says you have to have both. That would save lots of tree and free up a lot of digital space.

This author believes an organization’s people are its most important asset. Workforce management therefore is one of the most critical responsibilities given to those who manage people. The quality of managers given that responsibility and the adequacy of investment in their training will have a profound impact on the organization’s performance and its viability in the future. Training may be expensive but cost out the alternatives.


Time To Rethink Pay Structure Design?

In 2006 a research-intensive organization designs new pay structures for its employees. The Scientists & Engineers represent about 2/3 of the workforce, mostly data scientists and software engineers. The organization creates career ladders with four levels in each job family (Associate, S/E, Senior S/E and Principal S/E). A comprehensive review of prevailing practice and competitive pay levels produces four grades for each family and pay ranges with a midpoint to midpoint progression of 20% and ranges that are 50% wide. The framework is set for the administration of a merit pay system.

For the next five years prevailing pay rates in the relevant markets are flat, due to a significant economic downturn. Surveys show that pay- rates are going up at 2-3% per year and that structures are being increased by 1 – 2%. The organization was able to afford a budget of about 3% each year and believed it was tracking the market levels with that level of spending. And the pay ranges were increased by 1% each year. Since the organization was a government contractor it was obligated to pay no more than market average and diverging from that created a breach of contract. But an in-depth analysis of the surveys by one of the data scientists discovers that the methodology used by the surveying organizations is flawed and statistical anomalies have been created because of this. The surveys reporting on budgets failed to realize that when the economic picture darkened a number of organizations stopped reporting their data. After all, if you have frozen pay and/or reduced pay or staff levels why would you bother to report in the survey, so you could pay for data you would not be acting on? The surveys of pay levels suffered the same deficiency and they showed a continued increase in competitive pay levels when in fact that result was attributable to sample change and market levels had not actually moved.

Now the organization faces two problems. 

  1. One is that their spending level has outpaced the true market movement and since they were at market previously they were now ahead of the real market average.
  2. The second problem is that employee pay rates are still significantly below midpoint. Since their pay policy included a clause that stated that satisfactory performers should reach their range midpoint in about 5 years they clearly had not fulfilled that commitment. Employees began to complain that they must be paid well below competitive market rates because of this. So the reality was that the organization had run ahead of the market but was perceived to be below market. How could this happen if it had supposedly followed the market?

A 40 year analysis that I did recently shows that on average pay levels have moved about 1% per year faster than cost of living and about 1.5% per year more than average pay structure movement (see a prior posting). This means that an employee newly hired or just entering a job who starts at or near the minimum is going to take about 18 years to reach range midpoint. This is going to create discontent, since being fully competent in a job but being paid well below the market rate contradicts how pay levels should be distributed in pay structures.

Perhaps the realities of the last decade (and likely near future) makes the traditional pay structure design obsolete. There is greater pressure on organizations to motivate the most capable employees to contribute their best. Yet research has shown that when an outstanding performance rating results in a pay increase that is only 2-3% more than someone rated as satisfactory it is unlikely that they will view that difference as warranting the extra effort and as being equitable. And being below range midpoint for very long is likely to be viewed by exceptional performers as a failure on the part of the organization to appropriately reward their contribution.

The most common range width is 50%. So is a 50% range too wide for the current context? Should ranges be narrowed to 25% so those entering a job at the range minimum can be brought up to the midpoint in a more reasonable time? One approach is to have narrower ranges for jobs with shorter learning curves and with limited ability to contribute much more than satisfactory performance. And many organizations do that. But is 50% still workable even for professional and managerial jobs, or should they be truncated as well? The “broad-banding” fad that seems to break out every 7-10 years has pretty much withered again… 100-200% ranges are useless when it comes to administration, since they no longer serve as control parameters. When I bring the possibility of narrower ranges up I typically get resistance because there is a perception that a lot of headroom above competitive market levels is necessary to accommodate long service people. But how many jobs allow incumbents to contribute value at a level that warrants paying more than 20% over market? I get considerable resistance when executive jobs are being considered, but there has been a trend to use variable pay much more, to provide upside income opportunity that is warranted when results are stellar. Tech jobs have really created a challenge. When Bill Gates said his top programmer was at least ten times as productive as the second rank does that mean pay levels have to reflect this?

That would challenge organizations in two ways:

  1. justifying that relationship
  2. finding the money to fund it.

It is surprising that many tech firms have not turned to variable pay more extensively, since in theory extraordinary contributions could be compensated with one-time awards, avoiding rapid compounding of fixed cost payroll.

The existence of wide ranges when time-based step progression is used can produce excessive pay levels… and pay levels related to longevity rather than performance. This is still common practice in the public sector, although steps are disappearing at many entities. And since inflated base pay rates inflate generous pension benefits it results in indefensible benefits cost escalation. There is less pressure on costs in the not-for-profit arena and it is difficult to convince city councils, county boards and state legislatures to begin to exercise the necessary level of restraint, particularly near elections. If steps are to continue for some jobs it seems reasonable to limit pay ranges to no more than 20% in width.

Each organization should seriously question whether their current pay structures and pay administration programs are effective and appropriate in this era. Although it is difficult to predict the next 10-20 years most Economists would advise that organizations anticipate wide economic cycles with patterns that have not been seen in the past. In 2018 there is little evidence of increasing pressure on pay levels even while unemployment is at the lowest level for many years. Continuing to operate with pay structures and practices that seem to be out of synch with the environment seems a chancy strategy.

If I were running a start-up or emerging organization I would put tight constraints on base pay levels and I would look to well-designed incentive programs to motivate the desired results and to control the growth of fixed costs, such as base payroll. When downturns in fortunes materialize it is far easier to retain critical talent and avoid dysfunctional downsizing when you are not faced with cutting base pay to achieve immediate cost reductions. Every base pay adjustment is a career annuity – employees do not accept your request for a refund when revenues drop. They after all base their standard of living on base pay and look at reductions as a breach of contract – emotional even if not legal. Replacing a portion of base pay increases with variable pay awards is not a simple task however. An earlier post addressed a test of readiness for using variable pay. And there may be employee resistance. But an “all base pay” direct compensation package has all of the issues just discussed. And aligning rewards with organizational performance in addition to individual performance may create a shared destiny mindset, which should convince employees they are a part of a larger organization and the funds for rewards has to be generated somehow.

Evaluating Employee Performance: Challenging Necessity

In order to reward performance an organization needs to be able to define and measure performance. At the individual level performance is defined using criteria and standards… what needs to be accomplished and how well it needs to be done. The pre-requisites for an effective system are:

  1. performance is defined clearly and is understood by all involved parties at the start of the performance period,
  2. performance is continuously measured and feedback provided,
  3. performance is evaluated in a manner that is considered fair and appropriate by the employee.

Easy to say… not so easy to do effectively.

Performance standards should be set at equivalent levels of difficulty across the organization. Employees are very sensitive to equity and if standards are uneven across units or evaluators it will precipitate feelings of unfairness and lead to dissatisfaction. A fundamental question that must be answered is whether an employee’s performance will be measured against standards or against the performance of others. Although it is possible to develop descriptions of what constitutes “outstanding” or “fully meets standards”performance those standards must be interpreted by individuals. And individuals vary in their interpretations. When a trial run of a newly designed performance appraisal system was done at a major utility the head of Operations with over 600 subordinates rated 1% of his employees “outstanding.” The head of Public Relations rated 60% of his employees “outstanding.” It was very obvious that the two managers had interpreted the description of that level of performance differently. The difficulty of establishing reasonably equivalent interpretations of the rating scale levels is daunting but must be addressed if a system is to be accepted as valid.

It may be argued that ranking employees against peers avoids the difficulties associated with rating against a scale. The problem with ranking is that a considerable amount of variability in what employees in a unit do often exists. This increases the difficulty of comparing contributions. Another issue is that some employees may be long service and be more competent in all aspects of their job than those who have less time on the job. It is very difficult to adjust expectations based on the competence level of employees and some way must be found to accomplish that. In a large research organization I consulted with one technical executive had over 1700 scientists and engineers in his organization. He used a ranking system to evaluate performance on a relative basis. My original skepticism was overcome after I reviewed the process that was in place to do the ranking.

Each Supervisor would rank his or her subordinates, using a list of criteria that should be considered. Then each Manager put subordinate Supervisors in a room and conducted a “calibration session,” which consisted of reaching a consensus on a merged ranking. This process was completed all the way up to the executive and he ranked his direct reports. The system was viewed as fair by most employees but I felt one refinement should be made. The population consisted of new graduates (classified as Associate Engineers/Scientists), those who were competent in most or all aspects of their jobs (Engineers/Scientists), those who were fully competent and who took responsibility for leading and developing others (Senior Engineers/Scientists) and a few who were leaders in their fields and who were expected to create new knowledge (Principal Engineers/Scientists). All employees were classified into these levels based on their competence using a rigorous review process and expectations obviously differed between levels. The refinement was to rank employees against others classified at the same level in the career ladder, which would make it easier to adjust standards based on what reasonably could be expected of individuals.

This became even more critical when employees were arrayed on a histogram with current pay level being the vertical axis and ranking the horizontal axis. Here a second refinement was proposed… change the vertical axis to the compa-ratio. This change made it possible to determine what size of adjustment should be made to the current pay rate. A regression line was created to show how the current rate compared to the rate warranted by performance (rank in the population). Employees paid in the lower part of their range who had high rankings would receive larger increases than those paid in the upper part of their range and who had similar rankings. This is the equivalent to using merit guide charts that tie pay actions to both performance and position in the pay range.

Organizations who evaluate performance against standard definitions generally have scales with three t five levels. Some feel performance should be evaluated as either does not fully meet standards, fully meets standards or significantly exceeds standards. Using an “unacceptable” rating usually triggers a performance improvement plan and it is often considered a management failure if actions have not been taken during the year to correct below standard performance. And having an “outstanding” rating is often seen as difficult to differentiate from a “significantly exceeds” rating. There is not a clear right answer, thus the variability. In some jobs there is limited discretion and complexity and some organizations may use a binary “pass-fail” distinction for incumbents.

No matter the number of evaluation levels used there remains the challenge of comparing individual performance to a rating scale. By using case studies raters can be charged with rating hypothetical employees and then being asked to reconcile their ratings to those of a number of peers in training groups. Over time this can calibrate how raters view the levels used in the rating scale.

If done correctly, applying the fundamentals of sound evaluation, both a ranking and a rating against standard system can work. The current avalanche of articles in the practitioner literature on continuous evaluation and feedback is welcome. No one can manage performance by following an administrative procedure at the end of the performance period if every employee does not know what is expected, how they are doing, and how to get better… every minute of every day.

U.S. Leads All Highly Developed Countries… In Income Inequality

Personal opinions vary about how national income should be distributed and everyone is entitled to their own opinion. Economists, Political Scientists and Sociologists attempt to predicate their views on tangible evidence when they are practicing in their field, although they too are free to adopt their own opinions independent of the facts.

Joseph Stiglitz, Nobel Laurate in Economics, pointed out facts in his book The Price of Inequality, specifically:

“Recent U.S. income growth primarily occurs at the top 1 percent of the income distribution, those at the bottom and in the middle are actually worse off today than they were at the beginning of the century, and America has more income inequality than any other highly developed country.”

Since Stiglitz’s book was published the 2018 tax code revisions have been made and they are regressive, which means they will exacerbate inequality. Stiglitz then transitions into a more opinion-based mode: “In a democracy supposedly based on one person one vote, the 1 percent could be victorious in shaping policies toward its own interests.” He attributes the realities of this unequal distribution to low voter turnout, a system in which electoral success requires heavy investments, and the belief that those with money have made political investments that have reaped large rewards – often larger than the returns they have reaped on their other investments.  Since there is no sound scientific analysis that can be used to unequivocally establish a causal relationship between wealth and influence these observations must be treated as one person’s (albeit a highly qualified person) readings of the tea leaves.

Each organization must make decisions that allocate resources internally. Microeconomics has provided substantial theory grounded in research that provides policy makers with guidance about how to manage rewards. An organization that attempts to be a low-cost provider (i.e., Walmart) may choose to pay the lowest wages possible and to limit eligibility to employee benefits. An organization that is seemingly identical that uses different cost calculations may choose to pay more and provide more benefits (i.e., Costco). Walmart’s average pay has been 25-30% lower for non-management jobs than Costco’s, yet Walmart paid its CEO about 5 times what Costco did in 2014. The median income of a Costco customer is much higher than a Walmart customer, which would suggest the organization could charge higher prices, although they generally do not. Walmart is criticized by the government, the public, unions and many employees for their rewards strategy. But Costco is sometimes criticized by its shareholders for investing more in pay and benefits. There is no “right” answer to distributing income within the organization. Different parties at interest will have differing opinions about what policy should be. The Costco CEO defends the higher investment in people by comparing its turnover to that of its competitors and pricing that turnover, thereby justifying the policy driving compensation to a higher level. There certainly is a presumption by Costco management that higher pay has a positive impact on turnover.

The comparison of Walmart to Costco illustrates that income distribution within an organization is a complex issue, with many contributing factors impacting what results will be, given any policy. One of the metrics that has been used to attempt to address “fairness” of internal pay distribution is the relationship between CEO pay and the pay level of those at the bottom of the pay structure or the pay level of the average employee. The U.S. private sector multiples are dramatically higher than any other country at an aggregate level. But to decide whether this is a “fair” multiple or an optimal one the economic system must be considered, as well as the type of organizations included in the analysis. Any multiple is based on opinion and it is reasonable to assume it should be determined independently by each organization.

If the founder of a start-up becomes extremely wealthy because the person came up with an idea, turned it into a business plan and executed a strategy that created more value than the CEO of a Fortune 50 company did any comparison between the two must be tempered by an understanding of what drove the outcomes. If the start-up never made a profit and virtually all of the wealth was a result of a sky-high valuation of the organization, whether it was publicly traded or private, one might view the outcome as being the result of a bet on the future by investors. And investors make choices about when to buy and when to sell, since they are free to do so. If the Fortune 50 company CEO pay was the result of a Board decision to award a high salary and cash bonus it seems fair to judge the decision by asking a few questions: 1) did the performance of the organization produce a high return to owners?, and 2) did it result in high salaries and cash bonuses for the rest of the employees? If there was little correlation between organization performance and the rewards paid out the judgment of the Board should be questioned. And if the CEO and other highly paid employees did well at the expense of other employees someone must decide if this was justified.

Recent analysis seems to indicate that increasing income inequality is driven by certain types of organizations, industries or sectors. My consulting work in the public and not-for-profit sector has shown me that there tends to be much less difference between the top and the bottom pay ranges in the typical public sector pay structure than in the private sector. And since there is less usage of variable compensation and equity ownership the variation in total compensation is even greater in the private sector. For example, in public utilities, library districts and many cities and counties there is seldom more than a 10 – 15 times variance between the highest and the lowest paid employee. Since the most commonly cited CEO – average employee multiples in the private sector is 200 – 400 this is a very large difference between the sectors. The public sector does provide much richer benefits than the private sector, but the Economist Intelligence Unit has priced the gap at no more than 30 percent. The average large corporation provides its CEO with a salary of 1 million dollars or more, which drives this multiple to such a high number.

All of this leads to a conclusion… there is no “right” or “fair” multiple, at the organizational or national level. But there is a lot of emotion generated by numbers like a 400 multiple when reported in the press. Economists, Politicians and Sociologists generate tangible evidence regarding what exists. But their personal opinion about what is “right” or “optimal” is just that – an opinion. Where an individual stands in the income hierarchy will certainly influence opinion. And where one falls on the Liberal – Conservative political scale will also influence how one reacts to the current income distribution.

The federal government must influence national income distribution through taxation policy, and in some countries in Europe there is deliberation about capping executive income. CEO’s must deal with income distribution within their organizations. Individual citizens should form their own opinions and act on them… in how they vote, in how they feel about income distribution within the organization they work for and in how they express their views individually and collectively. Having the most unequal national income distribution among highly developed countries can be viewed as desirable or undesirable, depending on one’s perspective.

Extrinsic AND Intrinsic Rewards: NOT Extrinsic OR Intrinsic Rewards

There have been claims that extrinsic rewards can destroy intrinsic rewards. But these claims are supported only by lab research that lacks generalizability and is refuted by substantial field (aka real life) research that supports the opposite position. Yet the claims continue to show up periodically and they have wasted a lot of ink and speaking slots at conferences.

The need is for BOTH intrinsic and extrinsic rewards if employees are to be satisfied and motivated to do the right things. The most cited basic needs are

  • autonomy (being in control and having choices),
  • competence (being able to develop and use personal capabilities) and
  • relatedness (being a part of something significant).

And performance is a function of ability and motivation (a multiplicative formula). So if organizations are going to fulfill their missions and meet their objectives they must have the right people in the right roles doing the right things in the right ways for the right reasons. To put it simply they need motivated and satisfied people performing the work of the organization. That is more likely to happen if the people have their needs met. And their needs are more likely to be met if an organization creates well-designed roles.

A well-designed role sets the stage for both motivation and satisfaction. The research done on effective role design has provided the characteristics that must be present. Employees must be capable of doing what needs to be done, want to do what is needed, be allowed to do it and know what it is. And if they are to be motivated to put forth their best effort and to focus it on what the organization needs they need to feel they are rewarded in a manner that is fair, competitive and appropriate. If someone is to endure a thirty year career on an assembly line doing repetitive work they are going to expect adequate extrinsic rewards (aka, pay), to partially make up for poor role design and for the lack of intrinsic rewards. One characteristic of intrinsic rewards is that employees in effect have to give them to themselves, based on their own psychological makeup.

Someone deriving considerable satisfaction by doing things they value and in a way that provides the basic needs may place less emphasis on extrinsic rewards, since the intrinsic rewards may meet operative needs. And it is possible that they may derive satisfaction from receiving substantial extrinsic rewards (pay or recognition) since it signals their personal achievement.

But even if the amount of total rewards is adequate it is unlikely that an employee can be satisfied with only extrinsic or intrinsic rewards. The two are not related in a formulaic tradeoff function. That is, employees may be placed in roles that deliver a lot of both, a little of both or a mix of both. But there needs to be an adequate amount of both types of rewards if satisfaction is to manifest. What is puzzling to researchers is that subjects might seem satisfied and motivated even if there is little evidence of intrinsic rewards being available. Some people are upbeat and optimists and find coping strategies that minimize the unpleasant. Others cannot be satisfied even it they are showered with excessive amount of both types of rewards.

Some derive great satisfaction from doing error free work or out-producing everyone else even if they go home exhausted. Others derive satisfaction from figuring out how to do the least amount possible without being censured or terminated. Some will throw tennis balls at targets for long periods and enjoy the game. Others will find it pointless and view it is a punishment. As a result, researchers must adopt humble aspirations when setting out to support their hypotheses, unless they are capable of psychoanalyzing their subjects so they can adjust for inevitable differences between them. It is likely that the research study sample will include people with a variety of needs and taking an average of the group tells us little of value.

So let’s accept that those who receive acceptable amounts of both intrinsic and extrinsic rewards are more likely to be satisfied than those who do not. And let’s expect that organizations doing a better job of providing adequate amounts of both will have more effective workforces. It will open up a lot of print space and conference presentations so they can focus on important issues that really need to be addressed.

Creativity & Innovation: How Does An Organization Define It?

Winning in today’s highly competitive and dynamic environment requires being out front of the other organizations offering their value propositions to the same end users. If my software is better or cheaper and/or if I can get it to market before others I can enjoy premium pricing until they turn it into a commodity. But that means I have to have some advantage.

Often innovation is defined as invention… something that has never been now exists. And creativity is needed in order for invention to happen. But there are other forms of innovation. I can improve existing technology or products and that can generate value. Or I can integrate features that already exist but have not yet been combined into a single device. Jobs led the creation of the iPhone and just about everyone would credit him with innovation. But was it in the form of invention, improvement or integration? Was the magic of this product the ability to do so many things on one device? Did it actually do something that could not have been done with any device at that time? As sales skyrocketed not too many Apple folks were concerned about putting a specific label on what made the product possible. What was clear was that creative people innovated and product success was over the top.

Organizations increasingly believe that talent is their most important asset. Although machine learning and other forms of artificial intelligence can be used to solve problems or create a synthesis of existing things (features; ideas; devices) can a determination be made as to what will succeed in the marketplace using technology? Although focus groups and other forms of marketing research can discover what people think they want history has shown that the answer is evident only after the product is in the end users hand being used. So if creative people are the key to innovation will the kind of innovation needed dictate the kind of creative people needed?

Talent management is all about attracting, retaining and motivating the people required to do what the organization needs done. If what is needed is invention what type of talent is needed? If improvement is required does that suggest a different type of talent? And when integration is the objective is yet another type of person required? One way to identify talent needs is to begin with the desired end result and work backwards.

In most for-profit private sector organizations the criteria used to evaluate organizational performance can be expressed using financial, operational and customer measures. Using an instrumentation dashboard based on a balanced scorecard approach the organization can create a report card to evaluate success. So if an investment is made in research and development with the objective of creating an easier way to control software that would appeal to a broad consumer market, resulting in a large revenue stream, each R&D proposal can be evaluated by projecting its success in achieving that result. PARC (Palo Alto Research Center) did that… several times they created something unique. Creative people were innovating at a high rate. Yet because PARC was imbedded in Xerox and the broader organization was unable to commercialize any of the innovations and ended up giving them away the desired end result did not manifest. There certainly was a lot of talent doing a lot of creative things but the organization relied on people focused on innovation and not on selling something. Looking back on this failure what could the organization have done to produce a better result?

If creative people are placed in a pure R&D setting and asked to invent new things they will be motivated to do so. If invention is what is measured and rewarded it is likely to be what people focus on. But if invention is only part of what is required for meeting objectives should the talent pool also include creative people who know how to take an invention to market, whether they are imbedded in the R&D unit or become the internal customer of the invention at the corporate level? It seems so obvious that he full range of talent is needed to get the whole job done. But the Xerox/PARC case is not the only example of invention not leading to success. GM could not learn from NUMI or from Saturn, two successful innovations in the auto industry. How can sophisticated organizations not recognize that the journey from idea to sale must be completed?

Creativity and innovation are the keys to success in this competitive arena. That is not to discount the value of being able to execute strategies that remain the same. Being able to reliably execute an operational excellence strategy may not be accepted as creative by some. But competing based on price is the key to success for many organizations. And putting together a well-functioning global logistics chain can require a lot of creativity, even though it is more of the improvement and integration type. An operational excellence approach to competing needs a different talent management strategy than a product innovation approach. People are wanted who loath waste, drive out variation in processes and execute reliably. A product innovation approach needs different characteristics. So GM and Xerox probably would have realized value from their experiments if that had mapped out all the types of creativity required and ensured the talent pool contained people who could provide it all. Their staffing, development, performance management and rewards management strategies could have been customized to attract, retain and motivate all types of talent required. And the organizational culture could have been evaluated to determine whether it provided a context within which the full range of talent could have peacefully co-existed.

The Source Of Sherlock Holmes Success

Solving crimes requires both analysis and deduction, as does the decision-making process. Inductive is gathering evidence and analyzing it. Deductive is developing a hypothesis and then testing it with evidence.

“It is a capital mistake to theorize before having data. Insensibly one begins to trust facts to suit theories, instead of theories to suit facts.” – Sherlock Holmes

Sherlock formulated theories based on observable facts. This is consistent with the “big data analytics” approach so popular today. Accumulate a body of data and then look for relationships and patterns. But does that not leave one open to error? It has been said that sufficiently tortured data will confess to anything. And academic research begins with a hypothesis and then accumulates and analyzes data to test it. This is to prevent “fishing expeditions” where the researcher first finds correlations and then formulates a theory based on them.

In large databases one can frequently find correlations, but that may not establish causation or in the case of Sherlock “who done it.” Today’s data scientists use artificial intelligence and machine learning to find explanations for things. But do they explain reality or only identify data relationships? If A is correlated with B that may provide little guidance, since both may be caused by C. And if causation is being sought did A cause B or did B cause A? . Meaningless correlations have subjected many to ridicule for failing to provide any rationale for why they might exist. For example, it has been found that executive compensation levels are correlated with company performance, suggesting that paying more might get you more. But equally plausible is that successful organizations can afford to pay more without Board or shareholder rebellions. So what is the truth? Of what value is the correlation in guiding strategy?

Holmes concluded that Dr. Watson should consider acquiring a physician’s practice by using evidence gained through observation. He assumed that the worn steps on the stoop of a Physician’s office suggest an active practice that perhaps Dr. Watson should acquire. But that data alone could be misleading. If there were other tenants in the building they could have generated much of the traffic. And perhaps the wear and tear on the stairs was caused by dragging bodies out after failed treatment. So in order to correctly diagnose Sherlock should have used multiple regression analysis to identify all of the contributing factors and their relative impact on the result. Anyone having done multiple regression analysis lacking a computer can attest to the reality that it would take hours, if not days, to come up with such an explanatory model. And of course computers were not available at the time.

So were his instantaneous conclusions just good guesses or the result of extraordinary intuition? Was he the master of deduction? This is a bigger mystery than he was ever asked to solve.