Entrepreneurial Strategies, Part 1

In his book, Innovation and Entrepreneurship, Peter Drucker presented how innovation and entrepreneurship can be a purposeful and systematic discipline. That discipline is still as relevant to today’s business environment as when the book was published back in 1985. The book explains the challenges faced by many organizations and analyzes the opportunities which can be leveraged for success.

Drucker wrote that entrepreneurship requires two combined approaches:  entrepreneurial strategies and entrepreneurial management. Entrepreneurial management is practices and policies that live internally within the enterprise. Entrepreneurial strategies, on the other hand, are practices and policies required for working with the external element, the marketplace.

Drucker further believed that there are four important and distinct entrepreneurial strategies we should be aware of. These are:

  1. Being “Fustest with the Mostest”
  2. “Hitting Them Where They Ain’t”
  3. Finding and occupying a specialized “ecological niche”
  4. Changing the economic characteristics of a product, a market, or an industry.

These four strategies need not be mutually exclusive. A successful entrepreneur often combines two, sometimes even three elements, in one strategy.

Being “Fustest with the Mostest” (FwtM) is about establishing a leadership position. In this strategy, the entrepreneur aims at leadership. It always aims at creating a new industry, a new market or a quite different and highly unconventional process.

Being FwtM does not necessarily aim at creating a big business right away. In fact, not every FtwM strategy needs to aim at creating a big business, but it must always aim at creating a business that dominates its market.

Perhaps because FtwM must aim at creating something truly new, something truly different, nonexperts and outsiders seem to do as well as the experts, in fact, often better.

To use this strategy effectively, it requires thought and careful analysis. In fact, for this strategy to succeed at all, the innovation must be based on a careful and deliberate attempt to exploit one of the major opportunities for innovation.

Being FtwM requires an ambitious aim; otherwise, it is bound to fail. The strategy also must hit right on target or it misses altogether. Once launched, the FtwM strategy is difficult to adjust or to correct.

After the innovation has become a successful business, the strategy of FtwM demands substantial and continuing efforts to retain a leadership position. Otherwise, we have just created a market for competitors.

The entrepreneur must work even harder now to make his product or his process obsolete before a competitor can do it. Working on the successor to the successful product or process must start immediately, with the same concentration of effort and the same investment of resources that led to the initial success.

Being FtwM is the approach that many people consider the entrepreneurial strategy par excellence. Of all entrepreneurial strategies outlined previously, it is the greatest gamble. FtwM is also unforgiving, making no allowances for mistakes, and permitting no second chance. But if an entrepreneur is successful with FtwM, the strategy is highly rewarding.

Binary Classification Model for Truck APS Failure Detection Using Python Take 3

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: The purpose of this project is to construct a prediction model using various machine learning algorithms and to document the end-to-end steps using a template. The APS Failure at Scania Trucks dataset is a classic binary classification situation where we are trying to predict one of the two possible outcomes.

INTRODUCTION: The dataset consists of data collected from heavy Scania trucks in everyday usage. The system in focus is the Air Pressure system (APS) which generates pressurized air that is utilized in various functions in a truck, such as braking and gear changes. The dataset’s positive class consists of component failures for a specific component of the APS system. The negative class consists of trucks with failures for components not related to the APS. The data consists of a subset of all available data, selected by experts.

This dataset has many cells with missing values, so it is not practical to simply delete the rows with missing cells. This iteration of the project will produce a set of results by imputing the blank cells with the value of -1. We will compare the results from Take 1 and Take 2, where we imputed the blank cells with the zero and the mean value.

CONCLUSION: From the Take1 iteration, the baseline performance of the ten algorithms achieved an average accuracy of 98.8001%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3983%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2187%, which was slightly below the accuracy of the training data.

From the Take2 iteration, the baseline performance of the ten algorithms achieved an average accuracy of 98.8348%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3967%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2187%, which was slightly below the accuracy of the training data.

From the current iteration (Take3), the baseline performance of the ten algorithms achieved an average accuracy of 98.8003%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3967%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2312%, which was slightly below the accuracy of the training data.

For this iteration, imputing the missing cells with the -1 value improved the average performance of all models only slightly. For this project, the Random Forest ensemble algorithm yielded consistently top-notch training and validation results, which warrant the additional processing required by the algorithm.

Dataset Used: APS Failure at Scania Trucks Data Set

Dataset ML Model: Binary classification with numerical attributes

Dataset Reference: https://archive.ics.uci.edu/ml/datasets/APS+Failure+at+Scania+Trucks

The HTML formatted report can be found here on GitHub.

什麼會給最大化?

(從我的一個喜歡與尊敬的作家,賽斯 高汀

當一個組織成功的時候,這個組織的主導人會決定給什麼來最大化。有一些選擇的是:

  • 老闆的薪水
  • 向股東分配
  • 股票價格
  • 員工的工資
  • 對客戶產生積極的影響
  • 對文化產生積極的影響

因此,如果您是當地的有線電視公司,您可能會決定用利潤在客戶服務或更低的費率上做額外的投資,即使這些選擇也許不會使對長期或短期的盈利能有最大化。如果您是一家上市的公司,您可能會試圖通過利潤回購去炒作股票價格。或者,如果您是一家有使命的公司,您可能會用利潤去重新多加投資在這組織的任務上。

大多數人認為一個公司的唯一目的就是去最大化利潤。這是不可信的,也是一個危險的謊言。一個組織的存在是為了讓人們的生活過得更好,而不是相反。

社會企業家的精神是一個有用的概念,原因是它從一開始就宣布了那些事項是優先。先賺取一些足夠的利潤來增長,但最後還是將大部分的利潤用於為客戶及他人來服務。

您不必擁有一個多彩多姿的標籤來建立一個讓您能感到自豪的組織。你只需要決定你想要完成什麼,然後就去做。

Binary Classification Model for Truck APS Failure Detection Using Python Take 2

SUMMARY: The purpose of this project is to construct a prediction model using various machine learning algorithms and to document the end-to-end steps using a template. The APS Failure at Scania Trucks dataset is a classic binary classification situation where we are trying to predict one of the two possible outcomes.

INTRODUCTION: The dataset consists of data collected from heavy Scania trucks in everyday usage. The system in focus is the Air Pressure system (APS) which generates pressurized air that is utilized in various functions in a truck, such as braking and gear changes. The dataset’s positive class consists of component failures for a specific component of the APS system. The negative class consists of trucks with failures for components not related to the APS. The data consists of a subset of all available data, selected by experts.

This dataset has many cells with missing values, so it is not practical to simply delete the rows with missing cells. This iteration of the project will produce a set of results by imputing the blank cells with the mean value. We will compare the results from Take 1, where we imputed the blank cells with the value zero.

CONCLUSION: From the Take1 iteration, the baseline performance of the ten algorithms achieved an average accuracy of 98.8001%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3983%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2187%, which was slightly below the accuracy of the training data.

From the current iteration (Take2), the baseline performance of the ten algorithms achieved an average accuracy of 98.8348%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3967%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2187%, which was slightly below the accuracy of the training data.

For this iteration, imputing the missing cells with the mean value improved the average performance of all models slightly, but not so much for the Random Forest algorithm. For this project, the Random Forest ensemble algorithm yielded consistently top-notch training and validation results, which warrant the additional processing required by the algorithm.

Dataset Used: APS Failure at Scania Trucks Data Set

Dataset ML Model: Binary classification with numerical attributes

Dataset Reference: https://archive.ics.uci.edu/ml/datasets/APS+Failure+at+Scania+Trucks

The HTML formatted report can be found here on GitHub.

[https://github.com/daines-analytics/tabular-data-projects/tree/master/py-classification-truck-aps-failure-take2]

Tactics, Part 1

In the podcast series, Seth Godin’s Startup School, Seth Godin gave a guided tour to a group of highly-motivated early-stage entrepreneurs on some of the questions they will have to dig deep and ask themselves while they build up their business. Here are my takeaways from various topics discussed in the podcast episodes.

  • Starting businesses is like being an architect. We get to decide what we are going to build, what land to build it on, and what materials you will use to build it out. We also get to decide whether we are going to build it big, how to finance, and whether we are going to bootstrap our way through. It is completely up to us.
  • However, if the deck is stacked against the business in some ways, we would want to know that going in. We do not have to do the business, but it is up to us if we are going to do the project. Once we realize we have something of value to bring to the table, the way we architect it, the way we talk about it, and the way we make choices will have a dramatic impact on whether we can deliver our intended goods/services.
  • Often it is not about whether our work is worthy, it is about how do we architect it so that people who do not know you will get the point. The people who succeed at their work are those who figured out how to market their work and bringing their work to market. If we can figure out the part about shipping our work and interacting with the marketplace, everything else will take care of itself.
  • A business plan is writing down enough details to make a sales call. We do not need everything ready, just enough details to make a sales call. We can assert that our stuff is working and be ethical about not taking someone’s money until we have something to deliver. The point is that, until we figure out how to be in front of somebody and ask them for financial resources, references, or the equivalent, we do not have a business. Without a written-down business plan, we can use to make a sales call, all we have is just a dream.
  • The way that yellow pages became a multi-billion-dollar business was a great example of demonstrating value. By demonstrating value upfront, the yellow pages won store owners’ commitment. If we have our customer’s commitment to pay us, we are in great shape, and we do not even need to raise venture capital money.
  • How we price is up to us. The key thing should be that our customers are getting the value they expect. We also need to be mindful of this chasm between something is that is free and something that is not. The free vs. not-free distinction is important because paying ten dollars is not that different from paying one dollar once the customers are convinced about paying. That chasm between free and not-free is a gap we need to help our customers to get across. Our customers need a story that can tell them something that we do is worth paying for.

Distribution and Cultural Destiny

In his podcast, Akimbo, Seth Godin teaches us how to adopt a posture of possibility, change the culture, and choose to make a difference. Here are my takeaways from the episode.

  • There are two ways to make money from making media, and both are related. One is to sell advertising. To sell advertising, we need to figure out the distribution. Distribution helps us to get the word out and get people to read or watch or listen to what we make. Distribution determines the media, and the media determines the culture.
  • There are also three macro factors at work. The first one is the filter bubble coined by Eli Pariser. It is the idea that when you give people a choice, they tend to consume media that they agree with ignoring this stuff that they disagree with. The second factor is Chris Anderson’s long tail, that when you give people choices, they will take them. The third factor is the idea of curation, self-censorship, industry standards, and so on. When there was an industry that needed to deal with a tight distribution model, standards set in.
  • But once you get rid of that distribution bottleneck, there is less incentive for having self-censorship. Another word, when the distribution changes, the content that gets made also changes.
  • With the advance in technology, there are two forces at work. On the one hand, new forms of expression can exist because distribution has changed through Technology enabled rap and hip-hop to exist because musicians and artists could use technology to make music good enough to listen to at home or in a cheap studio. The other force is that there is an inherent self-curation self-censorship, and self-quality ratchet there goes into place when there is scarcity. That scarcity can be time, attention, or other valuable resources. The distribution caused the publishers to apply more curation.
  • As we go forward, we need to work with and balance these two forces. We have audiences who want to absorb certain contents contrasted with society’s desire for ideas and images that edify us. We must realize that the distribution channel is now all of us.
  • When we spread ideas, we are as responsible for that as the network executive was and as the bookstore owner was. The ideas that we share are the ideas that spread, and the ideas that spread are the ideas that win. When we choose to spread an idea that is corrosive, it takes us away from thoughtful interaction. We must accept the responsibility that we are the distribution now, and we need to own the outcome that comes with that.

Binary Classification Model for Truck APS Failure Detection Using Python Take 1

Template Credit: Adapted from a template made available by Dr. Jason Brownlee of Machine Learning Mastery.

SUMMARY: The purpose of this project is to construct a prediction model using various machine learning algorithms and to document the end-to-end steps using a template. The APS Failure at Scania Trucks dataset is a classic binary classification situation where we are trying to predict one of the two possible outcomes.

INTRODUCTION: The dataset consists of data collected from heavy Scania trucks in everyday usage. The system in focus is the Air Pressure system (APS) which generates pressurized air that is utilized in various functions in a truck, such as braking and gear changes. The dataset’s positive class consists of component failures for a specific component of the APS system. The negative class consists of trucks with failures for components not related to the APS. The data consists of a subset of all available data, selected by experts.

This dataset has many cells with missing values, so it is not practical to simply delete the rows with missing cells. This iteration of the project will produce a set of baseline results by imputing the blank cells with the value zero.

CONCLUSION: The baseline performance of the ten algorithms achieved an average accuracy of 98.8001%. The ensemble algorithms (Bagged CART, Random Forest, Extra Trees, AdaBoost, and Stochastic Gradient Boosting) all achieved the top accuracy scores after the first round of modeling. After a series of tuning trials, Random Forest turned in the top result using the training data. It achieved an average accuracy of 99.3983%. Using the optimized tuning parameter available, the Random Forest algorithm processed the validation dataset with an accuracy of 99.2187%, which was slightly below the accuracy of the training data.

For this project, the Random Forest ensemble algorithm yielded consistently top-notch training and validation results, which warrant the additional processing required by the algorithm.

Dataset Used: APS Failure at Scania Trucks Data Set

Dataset ML Model: Binary classification with numerical attributes

Dataset Reference: https://archive.ics.uci.edu/ml/datasets/APS+Failure+at+Scania+Trucks

The HTML formatted report can be found here on GitHub.

The Three Conditions of Innovation

In his book, Innovation and Entrepreneurship, Peter Drucker presented how innovation and entrepreneurship can be a purposeful and systematic discipline. That discipline is still as relevant to today’s business environment as when the book was published back in 1985. The book explains the challenges faced by many organizations and analyzes the opportunities which can be leveraged for success.

In addition to discussing the five do’s and three don’t’s, Drucker also laid out three conditions for innovation. The innovators must always keep these guiding principles in mind as they tackle their work.

The three conditions of innovation are:

  1. Innovation is work; it requires knowledge.

The work of innovation is hard work, and it often requires great ingenuity. Also, innovators have a laser-like focus and rarely work in more than one area. Drucker pointed out that innovation, at its core, is hard, focused, purposeful work making very great demands on diligence, on persistence, and on commitment. No amount of talent, ingenuity, or knowledge can make up for the lack of the diligence, persistence, and commitment.

  1. To succeed, innovators need to build on their strengths.

While successful innovators look at opportunities over a wide range, they always ask this critical question. “Which of these opportunities fits me, fits this company, puts to work what we (or I) are good at and have shown capacity for in performance?” It is important for innovators to build on her strengths because of the risks of innovation and the demanding premium on knowledge and performance capacity required.

Also, just like any other venture, innovators have a “temperamental fit” with their innovative effort. Because innovation is hard work, businesses or people rarely do not do well in something they do not like or respect. Innovative opportunity must be important to the innovators and make sense to them. Otherwise, they will not be willing to put in the persistent, hard, frustrating work that successful innovation always requires.

  1. Innovation is an effect on the economy and society.

Another word, innovations result in changes. It could be a change in people’s behavior or a process. A change in the process can affect how people work and produce something. As a result, innovation always must be close to the market, focused on the market, indeed market-driven.