There is no way around Estimation and Decision Making in business. Lean and Agile blend the art and science of estimation, estimation accuracy and decision making. This article will explore how to approach these concepts by combining several areas of best practices all into one context. A list of the major areas of influence can be found at the end of the article.
The main objective of a Business - is to play an equation game. That is, a business receives basic inputs and produces outputs for a profit. Simple enough ...
However, wouldn't it be great if it were that easy. Businesses that operate with a sense of ease and elegance fought hard with the details to get it that way. And, these details will be at the heart of good estimations vs poor estimations.
Here's an illustration to start things off:
Would you rather be 90% accurate for estimating a $100,000 project?
80% accurate for estimating a $10,000 project?
In other words, is it ok to just be 90% confident in an estimate? What role does the money play within your organization when estimates are issued? What if the percentages of success are the same? Would anything change in the rationale as to how the decision is made?
For example, when was the last time you were 90% confident in a business decision? How did it go? When was the last time you were 100% certain of a business outcome only to have the outcome surprise you? Never? Yeah, me too ...
Estimation can be hard on the mind when the details matter and when the stakes are high.
If businesses are designed to make profits, then people must make sound estimates and even better decisions for the business to be successful.
Does your company frequently make million dollar decisions?
What if the numbers were larger? $1,000,000, $10,000,000 or even $100,000,000? How does the sum impact how the decision is made? Is time a factor when making decisions? If so, does time count for anything? Is speed preferred over estimation accuracy?
So what do we make of this ... ?
What is the difference between being 90% certain of a $1,000,000 business decision?
Being 90% certain of $100,000,000 business decision?
Here is the difference:
- 90% certainty of $1,000,000 = $100,000 left to the game of chance.
- 90% certainty of $100M = $10,000,000 left on the table.
- 90% certainty of $1B = a whopping $100 Million in the balance.
Congratulations! If you read this far - you have just been introduced to the concept of how information has economic value.
In other words >> Information is Money -- Just like Time is Money.
So let's explore a bit of the obvious about businesses and sound decision making:
People make decisions.
People run companies.
People make estimations.
People create tools to help them with estimations.
People create processes to help them make better estimations to make better decisions.
A team is a collection of People and there are evidence based team dynamics at play that will heavily influence estimations and outcomes. Many times these Psychological Team Dynamics are completely missed or ignored in a business context as the culture of these businesses are simply not there yet. Unfortunately, poor team dynamics will negatively impact estimation best practices which will negatively impact Business Outcomes. Only the most elite organizations will dare to go to this level of evaluation, but to those that scoff I suppose I have to quote Deming:
So, Businesses can choose to ignore modern science around team dynamics, culture and estimation accuracy or they can choose not to.
For example, in Agile Scrum, teams are expected to frequently meet to exchange knowledge. As time progresses, the very best knowledge gets exposed repeatedly to the entire team. As a result, the team should reach a healthy equilibrium where only the best knowledge is being collectively leveraged to tackle the most complex and taxing problems. If this is actually achieved, then how can the business not benefit? How could the company not produce meaningful solutions to complex problems if the teams are sharing knowledge and problem solving together?
The Answer is: They must be motivated by something more powerful than generating world-class solutions. Perhaps the teams are being driven to meet unreasonable expectations and they are finding it easier to skip all the Agile Meetings for one reason or the other. This starts an undesirable cycle that System Thinking predicts - what will happen is that the estimations will not be Team Estimations because the teams are not even meeting in the first place. If that were not enough, Lean Principles predict bottlenecks will persist such as the one found in the book called the Phoenix Project. The bottlenecks will be found at the Knowledge Hoarders. By now, you know the drill - poor estimations yield poor business outcomes.
The point here is that People are your business and culture will drive team dynamics. Strong team dynamics should produce better estimates.
Let's turn our focus back to that big pile of cash - How much money, time or resources should a company invest to capture more information before making a decision? How much money should a company spend to save $100M in the Billion Dollar scenario where the company was only 90% certain of the outcome? What we just did was put a dollar amount on Uncertainty.
In Business - every estimate and every decision tied to the estimate is an exchange of Resources for some sort of Desired Outcome. Therefore, every decision can be thought of a Resource Allocation exchange.
Now, we are prepared to explore techniques as to how to issue better estimates and ultimately make better decisions in business terms.
1.) Information has economic value.
2.) Businesses seek to maximize favorable outcomes and minimize undesirable outcomes by making sound resource allocation decisions.
3.) Uncertainty can be measured and quantified in mathematical / business terms for every decision.
All decision makers are constantly bombarded with partial information and are required to make a decision with less than "perfect" information.
Therefore, here is a table to establish the value of Information for a million dollar business decision:
|Certainty||Dollar Amount||Uncertainty Expressed in $$$|
|100% Wild Guess||$1,000,000||$1,000,000|
A short lesson in the 3 to 5 Why's from Lean and Agile Principles
Overconfident or unaware decision makers may grow agitated with this subject as their decisions should never be questioned. There is a whole direction to investigate regarding the Psychological Profile of Apex Decision Makers, but that is not the point of this article - my reason for mentioning it has to do with the conclusions drawn for the lesson of the 3 to 5 Why's Principal associated with Lean Best Practices ... sooner or later the end of the 3 - 5 Why's exercise will lead back to specific policies which will eventually lead back to a decision maker. This phenomenon is not to difficult to grasp, but it is also not easy to resolve for obvious reasons. The boss is the boss. There is a good illustration in the book by Peter Senge called the Fifth Discipline. The book introduces the concept of Systems Thinking which is what the 3-5 Why's is all about.
The story goes like this - There is oil on the shop floor of a manufacturer. Why1 ? asks the observer. The maintenance manager replies, "That's because the fasteners do not correctly secure the oil reservoir." Why2? Because folks in purchasing were pressured to save money 18 months ago and found a deal on fasteners, but no one bothered to ask the folks in maintenance to double check the specifications before the order. Why3? Who gave the savings mandate to the Purchasing folks? A highly paid decision maker of course. Why number 3 goes right to a person..
For this reason, Doug Hubbard from Hubbard Research, recommends that Decision Makers should be Calibrated. Decision makers rarely have perfect information which is why it is important for the overall health of the business that these decision makers are skilled at estimation and uncertainty reduction techniques. Calibration training is one way to improve these skills.
Smart decision makers leverage every advantage to improve outcomes. Poor decision makers guess and hope which is the same or worse as flipping a coin.
The logic goes like this in a traditional business context - the more important the decision, then the higher up the ranks it ascends within the organization. The escalation process halts once the correct person arrives at a "good", best or even Optimal Decision. Even decisions made by committees require sound Estimations.
Conversely, Lean and Agile Principles seek to decentralize decision making even for very important decisions. Moreover, these decisions are to be tackled by the team right then and there when they are first encountered. The idea here is that it takes too much time to run the decision up the flagpole and wait for direction - in large organizations that could literally translate into weeks and even months which, of course, can translate into lost dollars at very large sums and missed opportunities. Lean and Agile Principles state clearly that those individuals closer to the problem are more likely to generate more accurate estimates - and these estimates are more likely to produce the very best solutions that will produce the best outcomes. This is why there is a double benefit for decentralized decision making.
Therefore, how do organizations strike the right balance between Centralized and Decentralized Decision Making? How does Estimation play a role in Decision Making? How does improved estimation impact Operations? How does improved estimation influence Operational Policies?
Remember --- Estimations come immediately before Decisions. Therefore, better estimations yield better business outcomes.
Now, we can start to see the clash and challenges associated with traditional Command and Control business structures that conflict with Lean and Agile Decision Making.
Let's observe a "simple" Data Driven Decision Guidance Spreadsheet Tool to set some context.
This image is from an excel file that you can download from this article. The file is a generic model of a "Decision Tree". This file was designed to introduce a set of decision making concepts and is not intended to be used for actual decision making. The idea is to convey that any decision can have many layers of logic supporting the outcome. In this example, the decision is Binary or a "Go / No Go" decision.
The context is set where this is the very top level of a decision space or problem space. That is, soon I will show you what is below this top-level. The best analogy is a "proverbial iceberg" where this level is located at the very top of the decision tree.
For organizations that adopt a data-driving culture, then the iceberg could actually be a some sophisticated decision tree tool such as a simulation tool or some flavor of a Bayesian Network.
For non-data driven cultures, then these Decision Pyramids are usually contained within the mind of the "expert" or highly paid decision maker - aka - The Boss.
The very elite decision makers or the very elite organizations do both. That is, they combine highly intelligent and highly trained individuals with the very best decision tools. I like to use a military analogy where the very elite fighter pilots are placed in the most bleeding edge technology. It would not be so good to place highly trained individuals in woefully outdated technology and expect extraordinary outcomes. Think: Top Gun pilot in a crop duster against an average pilot in a modern fighter jet. Here - my money's on the average pilot. The same principles apply to high stakes business decisions.
Now, returning to the excel file, we can image that the outcome is all that a Decision Maker is concerned about - and, in this specific example, the outcome is Green. This means that the decision maker will move forward with some decision without the need of exploring all of the other details that went into generating the outcome.
How and why did the Decision Maker arrive at this conclusion?
Why does the Decision Maker trust the logic of this Excel file?
In short, nobody wants their decisions to be overly inspected or questioned. Unfortunately, the business world requires that all outcomes should be evaluated - not to attack the decision maker, but to strive for Business and Operational Excellence.
All of this --- should work ---, but why do so many organizations struggle with the very basics regarding workflow management, system and process interoperability, dashboards, KPIs, Metrics, etc -- all around the context of Decision Making?
The clues point back to the original intent of all of these tools. The secret is also hidden in plain sight.
Before we move forward - let's inspect what is under the top layer of the excel file.
All that I am demonstrating here is that many decisions may seem to be straightforward, but as the stakes go up so too should the analysis.
The excel file contains 6 Nodes with 9 cells inside of each Node.
Each cell contains simple logic via a formula to help the Decision Network determine the ultimate outcome at the very top layer - which will be Green or Red.
There's a total of 3 logic layers with the top layer being the most important.
Again, this is a meaningless Decision Tree and the formulas and logic are not connected to any real world scenario(s). However, do not let that mislead you to believe that the excel file is completely useless. It has value in the sense that it could be easily modified to match a real business process.
The layers could be teams in an complex IT Organization or represent Product Lines inside of a Manufacturing Process.
The cells could be modified to contain real data instead of Randomly Generated data. Then, the IF Statements inside of the Cells can be changed to trigger outcomes that helps the decision maker set boundaries based upon legitimate business rules.
Perhaps modifying this excel file - just slightly - could serve as a thought starter or serve as a powerful communication tool to evaluate a complex business process which carries considerable Risk and Business Value for the company. ( Did you catch that? I specifically did not use the word Risk until now for very good reasons.)
Doug Hubbard defines Risk as the cost of being wrong x the chance of being wrong.
If you were to run the simulation in the excel file 10,000 times or even 100,000 times you will find the the outcomes are approximately:
- 85% Green and 15% Red
If you download the excel file you will find 3 buttons like so:
The file will let you experiment a bit. You can set the number of trials you want to run in the Decision Simulation. Too many times in business estimates are provided with a single outcome - like rolling a single die and getting a 3 - how absurd would to sound if our logic would conclude that this particular die only rolls threes? This is exactly why a lot of businesses fall victim to inferior plans. And, look how easy it is (relatively speaking) to create very robust spreadsheet models to help with very important business decisions.
Therefore, if the value of our decision inside of the spreadsheet was worth $1,000,000 to the company, then our Risk is equal to $150,000.
We could also state that the value of each percent improvement of Certainty is worth $10,000.
Put another way - the value of information is equivalent to $100,000 for every 10% improvement of Certainty.
I leave you with a short Summary and Definition for Estimation. The article took a broad journey around the landscape of Estimation and why that word is so critical to overall business performance. Additional posts will build upon many of the concepts introduced in this lengthy article.
Definition - Estimation - The act and art of making a decision in a world of constant changing information. The decision made in this context does not contain all desirable information or does not contain "perfect information". If we had perfect information, then an estimate would not be required at all.
Estimates should not be perfect. (cough ... as they are an estimate)
People that make estimates should not be held responsible to perfection or excessive precision based upon the estimate.
Estimates degrade over time in most real Business Scenarios. Estimates have a shelf life and their quality degrades as more time passes after the estimate has been issued.
The rate of decay of the estimate accelerates when new information enters the problem space after the estimate has been provided.
Lean and Agile practitioners should act with the information they have - at the time that they have it.
Link to Excel File:
Here is a short Glossary of Terms. I did not get into Optimization with this post, but I will build upon this article and will demonstrate how Simulations and Optimizations are related to Estimation and Decision Guidance.
1.) Measurement - the act of reducing uncertainty
2.) Uncertainty - the measurement of what is not known subject to the problem space. The measurement is usually expressed as a percentage.
3.) Certainty - 100% information or outcome is known subject to the problem space.
4.) Percent Certainty - a number assigned to the knowledge of the problem space in between the two extremes of 0 and 100%.
Zero percent - one extreme - Keep in mind that many times people overreact to new scenarios and believe that very little or nothing is known about the problem space - and this can be trained out of teams and individuals that react this way to new challenges or new information that change their traditional decision making guidelines.
100% Certainty - The other extreme - Unfortunately, the other extreme is dangerous - Certainty - Everything is known relative to the desired outcome and problem space.
5.) Decision - in business this usually is connected with a Resource Allocation Decision. Naturally, the more resources in the balance associated with the decision, then the more "important" the decision will become in the eyes of the Decision Maker(s).
6.) Risk - The pain and undesirable attributes of uncertainty which are mainly expressed in loss of Time, Resources or Money.
7.) Process - A set of steps designed to make the business more efficient. A good process is the best friend of the Decision Maker as a good process will inherently reduce risk and uncertainty. This means that a good process with inherently protect and even optimize the use of Time, Resources and Money.
8.) Engines of Profit - A set of steps or a set of higher order processes that combine to produce a set of outcomes that drive the business towards profitability.
9.) Apex Process - a process of all other processes
10.) Apex Metric - a metric of all other metrics
11.) Optimization - when precision matters as is the answer is "100% Optimal" subject to constraints. The optimization is seeking to maximize Resources (time, money, people or equipment).
12.) Heuristic - Many times this term is used synonymously with optimization in business. However, a heuristic can find a solution that is "good enough", but not necessarily the very best decision available in mathematical terms.
Areas of Influence for this post:
Theory of Constraints
Probability and Networks
Industrial Operational Psychology
Evidence-based Science and Medicine - Sadly, there are many accepted practices in business that are not evidence-based.
Books, Authors and other sources of influence:
1.) Doug Hubbard - his books and the material found on his website.
2.) The Fifth Discipline by Peter Senge
3.) The E-Myth Revisited by Michael Gerber
4.) Start up material - Steve Blank and Eric Ries - their books and blog articles. Includes the classics such The Lean Start-up and Business Model Canvas.
5.) Optimization as understood within the context of Tuomas Sandholm's work. Excel Solver can illustrate basic to moderate use cases. Optimization is often used as a generic term and many times people mean Heuristic or estimation when the term Optimization is used to describe a business process or business problem. Resource Allocation Problems and Decision Guidance Systems can benefit a business tremendously if the correct application of Optimization is used.
6.) The Phoenix Project by Gene Kim, Kevin Behr and George Spafford
7.) Moneyball - math and concepts
8.) Richard Thaler's work in Behavioral Economics
9.) Blackbird Health - emotions matter in the workplace especially when solving for team dynamics. This is a very new and emerging science (relatively speaking), but it is important to distinguish evidence based emotional science from management techniques currently found in the workplace.(bbh1.com)
10.) Mike Cohen and his solid Agile perspectives which are found in his books and blog articles.(mountaingoatsoftware.com)
11.) Lean / Agile concepts explored by Dean Leffingwell
12.) Jeff Hawkins - his book On Intelligence and his work at the company he founded called Numenta. The excel file comes from my tinkering of trying to comprehend Networks.
1.) Cash - https:// codeslayer2010.files.wordpress.com/2012/02/moneypicmilliondollars04.jpg
2.) Coin Flip - https://goo.gl/images/tyUcNs