Sai Kotari

Parametric analysis of Variance (ANOVA) is used to compare multiple populations and even subgroups of those populations, tells us if there are any statistical differences between the means of three or more independent groups. ANOVA is extension of t-test which allowed us to test two groups if there is difference in means. If you have multiple groups, you could use t-test multiple times but a better way of doing that is through ANOVA.
Anova is the ratio of Variability between and Variability within.
 
Various ways of comparing variances for more than 2 population:
One-way Anova/Single factor: One-way means the analysis of variance has one independent variable, Measures single factor from multiple sources. It helps to identify if there are significant differences between the means of independent variables.  Factorial ANOVA/Two-Way Anova- Covers ANOVA tests with two or more independent categorical variables. It allows us to “account for variation” for the ROW level due to other factors as each row will have its own mean and std deviation.  This allows greater focus on Group differences making it easier to detect differences. Through this we are attempting to minimize the Error variance by detecting variance. Two-way/Two-factor with replication – Measures 2 factors, but has multiple repetitions of each combinatio Two-way/Two-factor without replication/ Randomized block design - Measures 2 factors  
Total variance:
Total variance is ratio of between groups by within groups.  Larger the difference, the more likely the groups have different means and that we should reject the Null Hypothesis.
 
a)     Variation within each group
b)     Variation between the groups
 
Concept of confidence interval:
ANOVA tells us groups are different, but it does not tell us what the difference is.  Confidence interval helps us assess the location. Pairwise comparison will lead to Type I error by compounding the error.
The most important statistics in the analysis of variance table are the p-value (P), S, R2, and adjusted R2 values. Collectively, these values tell if means are significantly different from each other
P-value
-        If P is less than or equal to the alpha level, one or more means are significantly different.
-        If P is larger than the a-level, the means are not significantly different.
If the ANOVA results indicate that there are significant differences, we can look at the individual statistics and confidence intervals to learn more about the differences.
S, R2, and adjusted R2 are measurements to show how good the model fits the data.
 
There are multiple comparison methods:
-        Fisher’s LSD – This is used in conjunction with ANOVA. While ANOVA tells us if two of the groups are different from each other, Fisher’s LSD tests two specific groups against each other.
-        Tukey Method – For pairwise comparison after One Factor ANOVA
-        Scheffe Procedure – Similar to Tukey Method, this method is used when Null Hypothesis is rejected after ANOVA, this test is used to determine where the difference exists.
-        Bonferroni Correction – Method used to detect false positives
 
When might you use Anova?  A few scenarios as example when we might want to use Anova:
 
·        A group of patients trying three different therapies and to compare if one therapy is better than the others.
·        Three different types of drinks for a group of people, and compare the reaction time
·        Students from different colleges take the same exam, check if one college outperforms the other

Delphi Technique is a systematic consensus building method for a problem or future event by generating opinions and forecasts which involves structured interaction with a group of experts. Uses a highly engaged communication process in allowing a group of experts to participate in problem solving. This is used in combination or by itself for decision making mostly in organisations that are geographically diverse.
RAND corporation invented this technique to forecast the impact of technology on warfare.
 
Advantages:
·       This technique generates number of independent judgements without face to face meeting
·       It leverages resources across geographies and departments
·       Investment of time and cost related to face to face meetings is avoided
·       Experts can adjust their answer each round based on group response
·       Addressing three main problems with Focus groups and/or brainstorming –
o    dominant personalities,
o    Group pressure, ideas are overrun and subverted by the group,
o    noise as it’s difficult to keep focus group on topic
o    Sometimes conclusions and ideas deviate from the objective
 
Disadvantages:
·       As this process is time consuming, it cannot be used for situations that require speedy decision.
·       Expert opinion is a belief that may or may not be true
·       Does not develop the rich array of alternatives
·       Consensus does not always mean the correct answer, Common ‘everybody agrees’ so may be this is right. Even if everybody agrees, the solution is wrong
·       Internal validity or reliability is largely unknown
·       Lacks the advantages of live discussion – understanding and assessing each view point
 
Planning and considerations of Delphi Study
Due to lack of single methodology and lack of guidance, proper planning is critical. Following needs to be planned:
·       Understanding the need – What problem it solves and why do we need it
·       Design of the survey tool – as communication is survey based, selecting a proper tool that is easy and reliable is important. Followed by pilot test of the tool for its reliability and validity
·       Size of expert panel
·       Implication on lack of anonymity
·       Level of consensus – Definition and rules on how consensus will be arrived based on data/information from the survey.
·       Timeline – Usually takes 2 weeks between rounds, so consider a larger timeline for entire process
 
Delphi Workflow: Following steps involved in the process:
1.    Problem indications - A precise and comprehensive definition of the problem
2.    Select facilitator - A neutral person, familiar with subject, knows process of data collection and analysis
3.    Select experts - 15-20 experts who have knowledge and experience, capacity and willingness, sufficient time to participate
4.    Round 1 - Questionnaire - Open ended questions (4-5) for idea generation. Can be replaced with focus groups or face to face interview
5.    Qualitative analysis - Ideas collated, summarised, and grouped based on similar viewpoints
6.    Round 2 - Questionnaire - Closed ended questions created based on information from round 1. Ranked based on importance and agreement, and sent back to experts for feedback
7.    Consensus? Else, repeat round - Rules of consensus is prerequisite to initiative DT. Continue rounds till consensus is obtained. 70% response rate preferred.
8.    Conclusion & Inference - Rank statements reached consensus from high to low. Use descriptive techniques such as plots for Inference.
 
 
Before we look at uses of Delphi and Nominal Techniques, lets understand Nominal Group Technique in Brief:
 
This is a group brainstorming process that encourages participation from everyone. Effective way of pooled judgement or decisions in a group that meets face to face. This technique can generate many creative ideas. It allows every member of the group to express their ideas while minimising the influence of the other participants.
A common challenge of Nominal group technique is managing discussions. Discussion if not balanced, conversation may turn into argument. Facilitator should remember that the primary purpose is clarification and not to resolve difference of opinions
Advantages of Nominal Technique: 
·       Decision-making process is shorter
·       Participants know each other, they get opportunity for equal participation
·       Distractions avoided
·       Ideas are visible to all
Steps involved in Nominal Technique:
·       Generate: Objective shared and discussed to generate as many ideas as possible
·       Record: Members record their ideas and share
·       Discussion: Each recorded idea discussed to determine clarity and importance
·       Voting: Participants vote on ideas, and help prioritise
 
When do we use these techniques:
 
Use of Delphi Technique
Use of Nominal Technique
Recap of Delphi Technique: Expert ideas taken anonymously and passed around the group until a consensus is met. Participants are physically distant, no interaction between the experts, individual response is recorded, and consensus is arrived.
 
·      To get consensus in a large group where difference of opinion may not resolve through discussions
·       To shortlist priorities, usually for future events
·       Develop policy - When there is no clear-cut resolution of a given policy issue
·       To avoid investment of time and cost related to face to face meetings
·       If decision making involves people from all geographies, various departments
·       To build consensus for a particular group
·       Should only be used on absence of Analytical technique
·       When we have longer duration for decision making as this technique takes time
 
Recap of Nominal Technique: Allows each member to write down and present the idea to the group, ideas made visible to all. Participants get to know each other as they meet face to face. Ideas are ranked and consensus made.
 
·       To generate lot of ideas
·       To allow group thinking specially for introverts and extroverts - some groups are vocal, some think in silence
·       Where members do not participate in surveys
·       If group does not generate ideas (normally)
·       If some team members are new to the group
·       When you want to identify priorities and select few alternatives
·       When time available for consensus is short and needs immediate implementation
 
 
Conclusion:
 
As a lean six sigma consultant, we are always faced with problems that needs out of the box thinking, fresh perspective, innovative solutions. It’s important to use “Divergent” and “Convergent” thinking to do this. Divergent thinking allows generation of lot of potential ideas, solutions and possibilities which is commonly known as brainstorming. While convergent thinking involves evaluating those options and choosing more relevant one.
 
We must consider using combination of these two methods to develop new ideas and solutions. As using just one-way thinking could lead to unbalanced or biased decision. Hence combination of Delphi with Nominal technique would be ideal by leveraging its strengths, which also includes brainstorming to allow maximum innovative ideas with deeper understanding of its capabilities.

Parkinson’s Law: In 1955, British Historian and Author Cyril Northcote published an article that described and defined the natural tendency of officials to make more work for each other. He also mentioned common observation that work expands to fill the time available for its completion.  He used the example of an elderly lady writing a postcard to her niece. Since she has no other work, much obvious simple task takes up entire day. Though he was hinting inefficiency prevalent at bureaucratic level from his studies.
 
A task feels bigger and more difficult the more time you allot. This leads to procrastination and do less important things first, takes mental energy with stress and pressure. If a Project has 2 weeks to complete, its most likely that team will start to put it together in last 2 days. Last minute thing that we all have in projects. Quality of deliverable will be almost same even if you have more time in hand.
 
I would like to discuss its impact in various fields and tips to master and benefit from the law.
 
A. Parkinson law in personal productivity:
 
I find this critical as we can not manage any organisational goal without understanding its impact at individual level.  If you allow a week to complete for something that usually takes 3 hours, it will take a week. Task expands in complexity and takes longer to complete.
 
Common examples: Last day before holiday is always busy, chaotic, feeling exhausted, brain fog, multiple tabs open on screen, not able to focus, failed multitasking. For ex. Writing a book, Preparing for Exam
 
Set shorter deadlines -Intense attention for shorter period is often better than extended, lazy attention over a long period of time. Set up time block - Set a timer for start and end time, push yourself to get it done during this time. If you are competitive, then see this as your personal competition that you are trying to improve yourself on this. 100% undivided attention is must. Eliminate distractions. Keep actionable tasks – Activity breakdown to clear tasks with measurable time is key. Avoid broad tasks which have difficulty estimating time required to complete Rule of Five - Begin with 5 most imp things you must do for the day. Stay committed to 5 task list Prioritise - Order 5 imp things in priority. Don’t go to 2nd unless 1st one is complete.  
Understanding how our brain perceives time helps make most out of the day by setting shorter time limits which increases focus and making sure we are not wasting our time. Limit time wasting activities - scrolling through social media, reading headlines again and again, checking emails
 
B. Managing teams and Project deadlines:
 
Managing project deadlines is a common challenge as it involves a team to work together towards the goal in that limited time.  Team manager is always found busy, so much so that their most important job of managing the team gets neglected.  A few tips to improve productivity:
 
Ensure activity breakdown – Projects always start with detailed project plan, but key here would be to have daily realistic goals, clear ownership and follow up. Create tighter deadlines – Self-reflection, assess your abilities and team’s abilities and estimate accordingly, Identify team’s strengths and weaknesses. Set up ambitious deadlines based on how long it took last time, do we have any time constraints. Analyse team mix – Understand your team mix, introverts and extroverts have different traits and approach to the tasks assigned. Harmonize on the individual strengths and weaknesses to maximize outcome Stop working late - productivity per hour declines sharply if someone works more than 50 hours per week. Encourage team to finish on time. Use 80/20 rule to your advantage - Identify most imp elements, focus on what matters and eliminate non important activities (VA and NVA) review your work, identify more important elements, and drop the rest Track your time - Identify critical tasks and time taken to complete, ROI of the task. Team will find there are activities that does not bring any value Make a time limit rule - Compressed time, for ex. Daily tasks. Make a rule to complete daily tasks in set time Take frequent breaks - During your moments of focused productivity, check how long you can stay highly focused.  Apply “Pomodoro technique" – break down work into intervals and enjoy the break  
C. Bureaucracy and demand management:
 
Two of his findings indicate organisations have tendency to expand:
- Officials wants to multiple subordinates to avoid direct competition
- Officials make unnecessary work for each other
 
While this is not true everywhere but sounds familiar too i.e. without considerable increase in workload, we have layers of management that continue to grow. Usually companies start with flat hierarchy, as they grow, they hire subordinates, and a pyramid starts to grow. As the pyramid gets large, it gets expensive.  
Create awareness - Companies should be more aware and transparent of time spent on all kind of activities Maximise value – Through understanding of value added and non-value added activities Demand and capacity management – Review demand for its relevance, efficiency possibility. Simple example of resource back fill without assessment of workload Consistent pyramid reviews - Revisit organisation structure against the set strategy and goals, ROI of the role  
Conclusion:
 
We can get more done in less time using Parkinson’s law. This gives a sense of achievement, allows better time management, focusing on one activity at a time. Important in today’s time when we all work from home, dealing with conflicting priorities between home and office chores. Key to getting it done is Commitment of building daily structures, breakdown of activities and adhering to time blocks.

Blind spot analysis is means to identify and deal with incorrect and outdated assumptions that could potentially influence decision making. The term “blind spot analysis” was first coined by Michael Porter, well known for his research on competitive strategy and Porter’s Five Forces. He believed that in business, outdated ideas or strategies had the potential to stifle modern ideas and prevent from success.
 
A blind spot is also known as conventional wisdom is an area in the range of vision that one is unable to see or unable to understand.
A common example for drivers is the area that one cannot see while looking at rear view. One has to turn his/her head to see if there is a vehicle, now we have better solutions: split rear facing mirrors, sensors etc. Other example is psychological – This blind spot occurs when we allow our emotions and thoughts, mostly unconscious to influence our behaviours. We are wired to make cognitive shortcuts based on experience and assume outcome. This mostly sets us on autopilot.  
Blind spots manifests in several ways due to following cognitive biases:
Anchoring bias – Rely on pre-existing information no matter how reliable that information is. For ex. For ex. We check out a handbag for 4000, and then look at one that costs 800. Prone to see the 2nd one as cheap. Availability Heuristic bias – Assumptions based on information, news stories. These days social media is mis-educating people on many choices that they can make Bandwagon effect – Following the rest of the world. Common example – Stock market investment, Business meetings – If most people agree on an idea then most likely one would be quiet if he/she does not agree Choice Supportive Bias – Usually leaders defend own choice, ignore downside of their own bad decision, Confirmation Bias – Choice made based on confirmation that we already believe, or we already know. Ostrich Bias – Subconscious bias to ignore negative information. For ex. Looking at amount of work left, instead procrastinate hoping it will go away Outcome Bias – Evaluating performance only on outcome even if it was mere luck Overconfidence – Based on opinion or gut, mainly people make this mistake in stock market Placebo bias – Belief that it will help recover Survivors Bias – judging based on surviving information, for ex. Building in old city have strong foundation, but what we don’t see is how many buildings collapsed. We only see what’s still standing, (a few) Selective Perception – Perceive message owing to frame of reference Blind Spot Bias – One believes he/she is less bias than others  
Common examples at organisations that affects decision making process:
 
1) Organisational blind spots –  Usually resulted out of defence to unrealistic strategy or problems resulting from policy, goals, expectations. Organisational blind spots arise when leadership and/or management is unable to acknowledge outcome of unworkable strategies leading to splitting, blame and idealisation.
 
This results in distorted organisation overview with many many blind spots. A few factors to consider are:
·       Employees/managers bringing best version to the bosses, hiding the gaps/skills that they lack
·       Telling boss only what they want to hear instead of what needs to be said
·       Problems not brought up to leadership immediately, managers afraid of appearing to not be in control
·       Measures do not represent reality, and leaders making decision out of data and not feedback
 
2) Leadership Blind Spot - Leadership blind spots are the specific areas that a leader lacks attention to or does not acknowledge gap (this could be skill). Leaders can be successful in certain areas but could have a weakness that gets offset and never acknowledged. A progressive and good leader would self-analyses and take this as an opportunity to grow. Common blind spots we see in leaders:
·       Doing it alone – not involving honest feedback or discussions
·       Being unaware of the behavior exhibited
·       “I Know” attitude
·       More and more data driven and time spent in meetings instead of meeting real people and talking
·       Actively avoid uncomfortable conversations
·       Blaming and playing victim – commonly blaming predecessors
·       Surrounded by Inner circle – Lacks diverse thinkers
·       Impulsive triggers – decisions made of experience or perception
 
3) Competitive blind spots - This is critical as failure to analyse competition leads to wrong strategic decisions, slow response to competition, weakening in its ability to seize the opportunity and many more. A simplified framework for competitive blind spot below:
 
Industry
Misjudging industry boundaries and trends, critical technologies
Customers
Misjudging customer’s changing needs, customer segmentation and buyer behaviour
 
Competitors
Poor identification of competition & their strengths, not analysing competitors’ strategy, Advertising, branding, expansion, pricing
Organisation
Weak organisational culture and structure, lack of skills, brand positioning, failure to capitalise own strength
 
Performing blind spot analysis:
 
Common approach used by many organisations is to check the various ways that unconscious bias could manifest.
Benjamin Gilad, Psychologist and philosopher developed a systematic 3-step method:
Step 1: Refer to the previous strategic decisions, its context of the decision, various factors involved and system of solution agreed. Conduct Michael Porter’s 5 Force analysis to identify change drivers and if organization overlooked any important aspects. Step 2: Collect competitive intelligence on the target company. Sources may include annual reports, letters to shareholders, interview in the press, public appearances, industry meetings etc. Step 3: Compare result of Step 2 with analysis of Step 1, Any contradiction with the analysis on Step 1 is a potential blind spot. Top executives/leaders are smart, capable people yet exposed to several decision biases that comes with situation they are in. Objective analysis of blind spots could greatly help leaders in identify the potential blind spots and treat them.
 
A few ways to overcome blind spots is by cultivating
·       Diverse networks – Embrace diversity and inclusion for diverse opinions, grow inner circle to include others
·       Building deeper connection in the organisation with the employees and not get blinded with results
·       Feedback loop – encourage telling truth versus what’s more convenient
 
Self reflection helps identify where are our blind spots and then accordingly manage them.