I find myself inventing terms sometimes, and lately I’ve been using the term killer customer as a parallel to killer applications. Today I thought I’d share with you what I mean by it and why it’s a good term to have in the back of your mind when considering start-ups.
A killer application is that thing about your product that is so awesome that it just compels people to buy it. It’s that first application of the technology that proves its worth. For example last year I came to talk with a student at the University of Oslo who had worked with some other people to create a software that reduced lag in networking. It would only have to be installed on a server and it would use some previously unused capacity to send redundant information and thus reduce lag significantly. My first thought when I heard about this was: “Give it to me! I’ll make us all millionaires”. Unfortunately (for me) it turns out it was open source, and the code was already meant to be implemented into the Linux kernel. But that’s really besides the point, the point is that I could see a killer application at once. Online gaming. Everyone that has ever played a MMO over a bad connection with loads of packet loss will understand why this is a good idea.
With this in mind, it's easy to imagine a killer customer: World of Warcraft. If this technology hadn't been open source it’s likely that they would gladly pay a lot to have it implemented. And with Blizzard on board it would just be a matter of calling those other MMO games to stack up the other millions. When that market is saturated you could go after video streaming, stock market information, and so on. A great opportunity.
The killer customers, thus, are the obvious customers that will give you cash flow quickly. Sometimes killer customers are those that need your product badly, and that will be glad to pay for it. Other times it may be someone that agrees to let you use them as a reference. For example I once met up with a start-up in Houston that was a spin-out from a major oil company, the company had extremely low market risk because the oil company had committed to being its customer should it succeed in productizing their technology. Having killer customers reduce your market risk, and will give you a much easier time getting funding and getting people to trust that your company will succeed. You don’t need to call them killer customers of course, just remember that the first customer is extremely valuable!
If you liked this post or any other post feel free to click the “follow” button to the right to stay tuned to new posts when they appear. You can also follow me on Twitter as @vetleen.
Showing posts with label market analysis. Show all posts
Showing posts with label market analysis. Show all posts
Wednesday, 21 July 2010
Tuesday, 13 April 2010
How to deal with uncertainty - the maximize options approach
In a recent blog post I wrote that it is becoming increasingly difficult to forecast the future. As I imply, the problem is choice: Lower distances between people increases the ripple effect of individual choices, and technology together with consumerism increase the sets of choices, and options available to each choice to create an exponentially growing possible futures.
"Difficult to see. Always in motion is the future.” (Yoda)
Simply put I would claim that the uncertainty of the future is a function of the number of people that make choices, times the number of choices available, times the number of options available in each choice, times a coefficient for the impact of each choice. As all the factors that go in to the function have higher values now than they did earlier the uncertainty increases. What I don’t mention is how to cope with the increasing uncertainty; this is what I will address here.
Economists love choices, because they apply math to find the “right” choice. Consider that you’re in a TV game show, you get presented with two choices, a) 100% chance of $ 10 000 or b) 50 % chance of either $0 or $50 000. Any economist will tell you that the options are worth respectively $10 000 and $25 000, so you should choose b) (this is because given enough similar choices the average outcome would be those numbers). Now, if however, you chose b) and you lost and had to go home with $ 0, did you make the wrong choice? Many will say yes, because you lost all the money. These people think that a good choice is a choice that you would not change after you see the outcome; this seems to me like it is a bad definition. I think that you always have to judge a choice based on the information that was available at the time, thus it was still the right choice! If you are presented with a choice of this type, you should follow this procedure.
The problem, as the clever reader has already deduced, is that you rarely know the numbers. And this is exactly what forecasting has been occupied with since the beginning of the industrial era; how can we put numbers on stuff that we know very little about? There are tons of books on this subject, so I won’t go in to it, but rather utter my proposition on how to deal with uncertainty in choices in these days, and especially in the years to come. First, let me repeat myself, in instances where numbers are available or could be attained, follow the above procedure. This is for all those other instances.
I already professed my love for platform technologies, but I haven't explained why, so here goes. The more malleable a technology is, the more uses it can have, thus the more ways it can be successful should it fail in its intentional use. The more adaptable a technology is (Cēterīs paribus), the higher is the probability that its owners will find an application that is profitable. Consider a hypothetical example, two companies are specializing in medical technology, both companies have a development cycle of 15 years, and neither has any information about what its competitors are doing. One company is certain that it can make a cure for let’s say AIDS, and the other is certain that it can make a platform that will cure every bacterial infection known to man, but would only be able to market it for one use at a time. Assume that the technical challenge is equal, and that the main risk is that competitors beat them to market. Which company would you invest in? Since neither has any information about what their competitors are doing (the premise of uncertainty), I would invest in the bacterial platform company. The AIDS company has one chance to succeed, if someone else beats them to market they are dead, finito. If someone else makes the bacterial technology, the company can just change its marketing to target another disease, because the drug can target all diseases, but just market towards one at a time.
By the same principle, even if I don’t believe in global warming I believe in environmentalism, because an earth with rainforests have more options that an earth without them. A country with a highly educated workforce has more options than one without it. A company with several paths to market has more options than one with only one path to market. A technology with many uses has more options than one with limited use. A user friendly computer has more options than one that’s not. A diverse education gives more opportunities than one that’s specific. This principle is universal.
But, you may ask, how does this relate to forecasting the future? Well, the concept I’m trying to explain has two implications for you. Firstly, it means that if there are high uncertainty go with the option that leaves more options open. Secondly, when creating something, try to leave as many options open for as long as possible. If you don’t know how things are going to turn out right now, maybe you will have a better understanding in the future, thus closing doors prematurely is extremely dangerous and will become even more so in the future.
To wrap it up, here are some simple predictions made following this principle:
If you liked this post or any other post feel free to click the “follow” button to the right to stay tuned to new posts when they appear. You can also follow me on Twitter as @vetleen.
"Difficult to see. Always in motion is the future.” (Yoda)
Simply put I would claim that the uncertainty of the future is a function of the number of people that make choices, times the number of choices available, times the number of options available in each choice, times a coefficient for the impact of each choice. As all the factors that go in to the function have higher values now than they did earlier the uncertainty increases. What I don’t mention is how to cope with the increasing uncertainty; this is what I will address here.
Economists love choices, because they apply math to find the “right” choice. Consider that you’re in a TV game show, you get presented with two choices, a) 100% chance of $ 10 000 or b) 50 % chance of either $0 or $50 000. Any economist will tell you that the options are worth respectively $10 000 and $25 000, so you should choose b) (this is because given enough similar choices the average outcome would be those numbers). Now, if however, you chose b) and you lost and had to go home with $ 0, did you make the wrong choice? Many will say yes, because you lost all the money. These people think that a good choice is a choice that you would not change after you see the outcome; this seems to me like it is a bad definition. I think that you always have to judge a choice based on the information that was available at the time, thus it was still the right choice! If you are presented with a choice of this type, you should follow this procedure.
The problem, as the clever reader has already deduced, is that you rarely know the numbers. And this is exactly what forecasting has been occupied with since the beginning of the industrial era; how can we put numbers on stuff that we know very little about? There are tons of books on this subject, so I won’t go in to it, but rather utter my proposition on how to deal with uncertainty in choices in these days, and especially in the years to come. First, let me repeat myself, in instances where numbers are available or could be attained, follow the above procedure. This is for all those other instances.
I already professed my love for platform technologies, but I haven't explained why, so here goes. The more malleable a technology is, the more uses it can have, thus the more ways it can be successful should it fail in its intentional use. The more adaptable a technology is (Cēterīs paribus), the higher is the probability that its owners will find an application that is profitable. Consider a hypothetical example, two companies are specializing in medical technology, both companies have a development cycle of 15 years, and neither has any information about what its competitors are doing. One company is certain that it can make a cure for let’s say AIDS, and the other is certain that it can make a platform that will cure every bacterial infection known to man, but would only be able to market it for one use at a time. Assume that the technical challenge is equal, and that the main risk is that competitors beat them to market. Which company would you invest in? Since neither has any information about what their competitors are doing (the premise of uncertainty), I would invest in the bacterial platform company. The AIDS company has one chance to succeed, if someone else beats them to market they are dead, finito. If someone else makes the bacterial technology, the company can just change its marketing to target another disease, because the drug can target all diseases, but just market towards one at a time.
By the same principle, even if I don’t believe in global warming I believe in environmentalism, because an earth with rainforests have more options that an earth without them. A country with a highly educated workforce has more options than one without it. A company with several paths to market has more options than one with only one path to market. A technology with many uses has more options than one with limited use. A user friendly computer has more options than one that’s not. A diverse education gives more opportunities than one that’s specific. This principle is universal.
But, you may ask, how does this relate to forecasting the future? Well, the concept I’m trying to explain has two implications for you. Firstly, it means that if there are high uncertainty go with the option that leaves more options open. Secondly, when creating something, try to leave as many options open for as long as possible. If you don’t know how things are going to turn out right now, maybe you will have a better understanding in the future, thus closing doors prematurely is extremely dangerous and will become even more so in the future.
To wrap it up, here are some simple predictions made following this principle:
- Mobile phones that allows anyone to create uses will have more uses than a mobile phone that don’t, thus cell phone producers that have open platforms will outlive those that don’t.
- Countries that have an adaptable workforce will be less affected by upheaval, because they can shift the workforce over in other industries temporarily or permanently should disaster strike in a specific industry.
- Renewable energy producers that use existing infrastructure, such as oil from algae, will be more successful than those that gamble on technologies like hydrogen that requires major rebuilds of gas stations etc, because they have more potential customers, quicker.
- On demand television will outcompete fixed programming, because people will have more options on when and where to watch. On demand television also have more options on how to make money - the business model.
- The deck of cards will survive Monopoly, because you can play many more games with a deck of cards. Also you can do magic, tell someone’s fortune or even use them for a raffle.
If you liked this post or any other post feel free to click the “follow” button to the right to stay tuned to new posts when they appear. You can also follow me on Twitter as @vetleen.
Monday, 1 February 2010
Familiarize yourself with the market in 3 steps!
Starting something new often requires you to read up on an industry and do those tedious swots, pestels and other analysis. The problem with these is often that the assumptions you take into the process are also the ones that comes out. If you think many small companies in the industry is an opportunity for intruders, well than that’s where you place it in the analysis, and that’s what you’re going to base your recommendations on. Only now it’s no longer just intuition, it’s derived from a model. Anyone else see the problem here? This framework doesn’t solve this, but helps you ground your understanding of an industry’s players in reality, and that’s a good start for a market analysis. The purpose of this analysis is the give you, the analyst, a tacit understanding of the industry as well as some models that describe it. So let’s dive in!
1. Define area, get lists
In my thesis I needed to understand “biotech companies in Norway”, I went to the business register and looked through the classifications that companies could choose from when they register the company. From this I chose (by using my preconceived notions) some classifications that may or may not contain biotech firms. The classifications included everything from "Research and development activities within biotechnology" to "Production of medical and tooth-technical instruments", as you can imagine some categories proved more relevant than others, but I’ll get back to that. Now I had a list of about 2000 companies, most probably not relevant since I included most categories that could be biotechnology.
This list of course can be anything; maybe you don’t want to look into a specific industry, but maybe a geographical area. If you want to start a company in your town, you could get a list of all companies in that town (however your dataset may be too big to actually get through) to understand what the major industries are and so forth.
2. Google, classify, sort, iterate
When I had the list, the next part was easy, I picked a category that I really believed in, "Research and development activities within biotechnology", started at the top and typed in each name in Google. The companies that have websites was easy, I just found somewhere it said in simple terms what they were doing and pasted it in next to the company name. For example the company Biosergen in Trondheim claims they are “a biotech company developing new drugs based on cutting edge biosynthetic engineering of natural products, combined with chemical synthesis”. Next to the “what we do” column I start making categories. This is where you should leave your assumptions at the door. Take the first two companies, and ask yourself “if I had to place both companies in a category what would that be”, for me it was “Biomaterials”, the third company then either fit’s in the category or needs a new category. After a few companies I realized this was not a good classification system, but by then I started understanding better, so I changed the categories. This is the iterate part of the process, change the categories.
The companies that do not have a website you can call and ask politely what it is they do, “Hey, I noticed you were classified as a --- company, but I can’t find your website. I was just wondering what field your company is in”. I noticed that I could do about 200 companies a day, and trust me, after a day or two you know the categories by heart and really gain an understanding of the industry that no SWOT or PESTEL analysis can give you. And if going through 2000 companies sounds like a lot, well, you are right, but over half of them were removed from the analysis because after 10 companies in the category I realized that I probably wouldn’t get any biotech companies from there. This of course depends on how thorough you need to be. If you have little time the try to spend a day, start with the most relevant category and just sample the rest, you still get a pretty good understanding of it all.
3. Model it
When you have worked for a while and are comfortable with your categories you should take a brake and define the categories you have come up with, then you start to see that some categories are overlapping or for some other reason should be better defined. For example for a while I operated with a category “Drugs” and a category “Therapy”, the distinction was really useless and I merged them, together with another category into “Treatment”. Other categories I ended up with was “diagnostics”, “research institution”, “services” and “supplements”. You should take the categories, define them properly, for instance “Treatment: Companies in this category are companies that a) at present or in the future will have as a core competency to treat ailment or b) facilitate that caregivers (such as doctors) use their products in order to give treatment or c) in any other way be a service provider based on their product that treats patients”. It is worth noting here that for my purpose I removed companies that weren’t based on biotechnology, because I was only really interested in that sector, not in for example hospitals.
Once you have the excel sheet with all the companies, sorted by category and with a short description of the company you can start modeling it. Make a box for each category, write down how many are in it, what they typically do, how big is this category, who are the biggest actors. Draw arrows and new boxes; try to make a model of the industry or area based on your analysis.
Extra tips and tricks:
If you liked this post or any other post feel free to click the “follow” button to the right to stay tuned to new posts when they appear. You can also follow me on Twitter as @vetleen.
1. Define area, get lists
In my thesis I needed to understand “biotech companies in Norway”, I went to the business register and looked through the classifications that companies could choose from when they register the company. From this I chose (by using my preconceived notions) some classifications that may or may not contain biotech firms. The classifications included everything from "Research and development activities within biotechnology" to "Production of medical and tooth-technical instruments", as you can imagine some categories proved more relevant than others, but I’ll get back to that. Now I had a list of about 2000 companies, most probably not relevant since I included most categories that could be biotechnology.
This list of course can be anything; maybe you don’t want to look into a specific industry, but maybe a geographical area. If you want to start a company in your town, you could get a list of all companies in that town (however your dataset may be too big to actually get through) to understand what the major industries are and so forth.
2. Google, classify, sort, iterate
When I had the list, the next part was easy, I picked a category that I really believed in, "Research and development activities within biotechnology", started at the top and typed in each name in Google. The companies that have websites was easy, I just found somewhere it said in simple terms what they were doing and pasted it in next to the company name. For example the company Biosergen in Trondheim claims they are “a biotech company developing new drugs based on cutting edge biosynthetic engineering of natural products, combined with chemical synthesis”. Next to the “what we do” column I start making categories. This is where you should leave your assumptions at the door. Take the first two companies, and ask yourself “if I had to place both companies in a category what would that be”, for me it was “Biomaterials”, the third company then either fit’s in the category or needs a new category. After a few companies I realized this was not a good classification system, but by then I started understanding better, so I changed the categories. This is the iterate part of the process, change the categories.
The companies that do not have a website you can call and ask politely what it is they do, “Hey, I noticed you were classified as a --- company, but I can’t find your website. I was just wondering what field your company is in”. I noticed that I could do about 200 companies a day, and trust me, after a day or two you know the categories by heart and really gain an understanding of the industry that no SWOT or PESTEL analysis can give you. And if going through 2000 companies sounds like a lot, well, you are right, but over half of them were removed from the analysis because after 10 companies in the category I realized that I probably wouldn’t get any biotech companies from there. This of course depends on how thorough you need to be. If you have little time the try to spend a day, start with the most relevant category and just sample the rest, you still get a pretty good understanding of it all.
3. Model it
When you have worked for a while and are comfortable with your categories you should take a brake and define the categories you have come up with, then you start to see that some categories are overlapping or for some other reason should be better defined. For example for a while I operated with a category “Drugs” and a category “Therapy”, the distinction was really useless and I merged them, together with another category into “Treatment”. Other categories I ended up with was “diagnostics”, “research institution”, “services” and “supplements”. You should take the categories, define them properly, for instance “Treatment: Companies in this category are companies that a) at present or in the future will have as a core competency to treat ailment or b) facilitate that caregivers (such as doctors) use their products in order to give treatment or c) in any other way be a service provider based on their product that treats patients”. It is worth noting here that for my purpose I removed companies that weren’t based on biotechnology, because I was only really interested in that sector, not in for example hospitals.
Once you have the excel sheet with all the companies, sorted by category and with a short description of the company you can start modeling it. Make a box for each category, write down how many are in it, what they typically do, how big is this category, who are the biggest actors. Draw arrows and new boxes; try to make a model of the industry or area based on your analysis.
Extra tips and tricks:
- You can use several categories for each company, but from different sets. For example a company can be "Treatment" in one column and "Oslo" in another. But in each set of categories you need to be mutually exclusive, if one category is "Treatment" and one is "diagnostics", you don't want a company to be in both. Have mutually exclusive categories in each category "set", but feel free to use more sets based on different criteria.
- The initial list should be made thinking it's better to include companies that's not what I'm looking for, than to exclude companies that I am looking for.
- Many lists that you can get also contains more information that the name of the company that can be useful when modeling, such as location, revenue, year founded an d so on.
- If you are more people, try spending a few hours on your own classifying the first 100 companies, meet and see what categories you have come up with. Why do you have different categories? Which are best? Try then to classify the next 100 together using a system you have agreed on. If you still have companies to classify separate and do different companies, you now have a common understanding of the industry.
- Use this analysis before the swot or pestel or whatnot, the findings here can be valuable in understanding what it is you will do, and what topics you should look into in the other analysis.
- After modeling try to talk to people who work in the industry, how do they perceive it, do they agree with your understanding of the industry? Did you miss anything?
If you liked this post or any other post feel free to click the “follow” button to the right to stay tuned to new posts when they appear. You can also follow me on Twitter as @vetleen.
Subscribe to:
Posts (Atom)