TL;DR
After the initial Lean Canvas it is easy to get lost on what to do next when you have so many assumptions so I use some techniques to help me keep on track.
I will take you through my train of thought with BARconWIFI.com as an example.
I will take you through my train of thought with BARconWIFI.com as an example.
Picking up where we left it...
In my previous post I talked about how I use the Lean Canvas in order to analyse and better visualise what I know or don't know about how I think my business model will work. But, a problem for me, specially at the beginning, is determining what to do next.
The thing is, once I have done the canvas, I am faced with so much uncertainty and possibilities that deciding what to do next is not trivial. Maybe you have had the same problem, where you find yourself asking: What should I go for next? Better yet: What am I trying to accomplish with it so that I learn the most out of it?
The thing is, once I have done the canvas, I am faced with so much uncertainty and possibilities that deciding what to do next is not trivial. Maybe you have had the same problem, where you find yourself asking: What should I go for next? Better yet: What am I trying to accomplish with it so that I learn the most out of it?
Well, Lean Startup, is about constantly performing a ruthless exercise on prioritisation and waste prevention. Taking the riskiest core assumptions, and learning as much as I can about them, as cheap as I can. That way I try to diminish the uncertainty about the whole business but at the same time move forward. This is what the whole MVP strategy is about.
Like Eric Ries likes to say, you are taking market risk (is someone going to buy my stuff) and turning it into technical risk (will I be able to build other stuff to test that and learn from it the answer?)
At the core of all this is the Build-Measure-Learn cycle. You build and experiment (or MVP) to test the market assumptions you are making. You then measure the response, and finally based on the data you can decide what you want to explore or build next. So... in order to start learning we need to start building.
Now, selecting what you want to build, requires you to have some idea of what you want to learn about. This is where the phrase that is often used to describe why startups fail comes in.
Paraphrasing Steve Blank: Most startups don't fail because they didn't build their products, they fail because they didn't build a product people wanted in the first place. So, the biggest risk one faces when starting a startup, is actually building the wrong thing and that is what the first step in the customer development process is about: customer discovery.
In customer discovery we are trying to find out if we are building the right thing. We want to validate that we know who our customers are, that we know which problem they have and that we are able to, not only give them a solution for that problem, but also that we are able to explain to them what that solution is, in an effective way.
However with so much information out there where do you go first?
Here is what I do to find out and organize that first wave of thoughts. I'm hoping it will be helpful to you.
Paraphrasing Steve Blank: Most startups don't fail because they didn't build their products, they fail because they didn't build a product people wanted in the first place. So, the biggest risk one faces when starting a startup, is actually building the wrong thing and that is what the first step in the customer development process is about: customer discovery.
In customer discovery we are trying to find out if we are building the right thing. We want to validate that we know who our customers are, that we know which problem they have and that we are able to, not only give them a solution for that problem, but also that we are able to explain to them what that solution is, in an effective way.
However with so much information out there where do you go first?
Here is what I do to find out and organize that first wave of thoughts. I'm hoping it will be helpful to you.
Down the rabbit hole.
Like I mentioned at the end of the last article I not only filled in the boxes for the BARconWIFI.com's Lean Canvas, but also had a section for questions that would come up as I was thinking and filling in each section of the canvas. This is definitively not by chance.
While I am filling in each box, I ask myself the questions that someone who would have to do the job, would ask me, or questions that would help me understand the way the business would work.
For instance, if I say that my customers are freelancers and I know I will want to communicate with them some natural questions follow. For example for customers:
For instance, if I say that my customers are freelancers and I know I will want to communicate with them some natural questions follow. For example for customers:
- Are these really my customers?
(Remember, biggest risk, not building the right thing for these people) - What similarities do they share?
(Everyone is different, but there needs to be a pattern if you plan on selling to them in a consistent manner) - Which ones are going to be my early adopters?
- How many of them are there? (Market size)
- ...
In the case of the channels. If I say that I assume will use Twitter to deliver some of the value and to communicate with them, then:
- Are they even on Twitter?
- How will I get them to listen?
- How will I get them to spread the word?
- ...
- Do my customers have this problem?
- How often do they have it?
- How painful is it?
- ...
There is a lot of stuff I don't know and that is exciting!
Prioritising
To prioritize I first write down the assumptions one by one on post-it notes. Then I distribute them in two axes.
The horizontal axis with how much would it take to execute the experiment. This is based on my own notion of what "risk" is. For me its a combination of the cost in time and money I think I would have to invest.
As an example, a round of interviews would be easier to do than let's say, build a twitter account with 100 followers in two weeks, or write down piece of software. Also, some experiments are easier to run if you have money to throw at them. So some times you could have the same experiment rank differently on the same axis depending on your resources (time, money, team, software...).
The vertical axis represents how much I think I would learn. Again based on my own notion of "learning". Again you will have your own set of learning priorities, but for me at this stage knowing that my customers really have this problem is more important that knowing if twitter is the channel to reach them, or if they will use my yet incipient solution the way I think they will.
The horizontal axis with how much would it take to execute the experiment. This is based on my own notion of what "risk" is. For me its a combination of the cost in time and money I think I would have to invest.
As an example, a round of interviews would be easier to do than let's say, build a twitter account with 100 followers in two weeks, or write down piece of software. Also, some experiments are easier to run if you have money to throw at them. So some times you could have the same experiment rank differently on the same axis depending on your resources (time, money, team, software...).
The vertical axis represents how much I think I would learn. Again based on my own notion of "learning". Again you will have your own set of learning priorities, but for me at this stage knowing that my customers really have this problem is more important that knowing if twitter is the channel to reach them, or if they will use my yet incipient solution the way I think they will.
By doing this I am already forcing me to think about how I would test each assumption and I write down the ideas for tests also. Once I do this, I usually start with the ones that that give me the most learning and at the same time are easier to test.
Again, you can use whatever method to prioritise y usually assign numbers and do a quotient. From here on I start taking them and clustering them around the experiments that I think would validate them. With that prioritised list in mind you can now chose which experiments to run.
But before running the experiment, I rephrase the questions in the experiment into testable hypotheses. By that I mean, you can determine if its true or false. Otherwise you don't know what it is that you are testing or if you found it or not.
For instance the questions: Are these really my customers? Is not clear enough for it to be testable.
In an effort to make it testable, I would change it to: Do they have the problem or problems I am describing? Ok that's better, but still not specific enough for me. I ended up with: I know they are my customers if they have had the problem in the past 4 months.
Now that we have a specific assumption that we are making, I set an acceptance criteria. You know, the line in the sand. Where in the sand? That, again, depends.
Trevor Owens from the lean startup machine describes it in his explanation of the validation board as the minimum criteria that needs to hold in order for you to be satisfied that the assumption (hypothesis) is true.
Since we really don't have any idea of what we are going to find in most cases, this is usually an educated guess. I like not to be too optimistic. If I do better than random, I'm fine with that. To me it means I am on to something.
Now the experiment I selected for this question is a set of problem interviews. The acceptance criteria I chose is that at least 6 out of 10 people that I interview say that they have had this problem in the last 4 months.
Once you have done that, you go and execute the experiment. Of course, you can run more than one experiment at the same time if you wish.
Ok, so based on the assumptions I am making I decided to run a couple of experiments in parallel. If you have some comments I would love to hear them.
What I chose to do was:
Problem Interview: Easy to do when your customers base is so approachable, and I know some people who fit the profiles. It yields the greatest learning for this part.
Landing Page stating UVP and sign up form to start collecting leads: Easy to do since I am using instapage.com (free) and lets you customise your landing easily as well as connect it to Google Analytics. You can read my other post if you want to know about other free tools I am using to build BARconWIFI.com.
Data Research of various kinds:
This is all due to the fact that I am already paying for socialbro.com which allows me to easily run this experiments and research on twitter without much hassle and that I know how to use some other free tools to automate most of the process.
If that were not the case, I am not sure I would have gone this way.
For instance the questions: Are these really my customers? Is not clear enough for it to be testable.
In an effort to make it testable, I would change it to: Do they have the problem or problems I am describing? Ok that's better, but still not specific enough for me. I ended up with: I know they are my customers if they have had the problem in the past 4 months.
Now that we have a specific assumption that we are making, I set an acceptance criteria. You know, the line in the sand. Where in the sand? That, again, depends.
Trevor Owens from the lean startup machine describes it in his explanation of the validation board as the minimum criteria that needs to hold in order for you to be satisfied that the assumption (hypothesis) is true.
Since we really don't have any idea of what we are going to find in most cases, this is usually an educated guess. I like not to be too optimistic. If I do better than random, I'm fine with that. To me it means I am on to something.
Now the experiment I selected for this question is a set of problem interviews. The acceptance criteria I chose is that at least 6 out of 10 people that I interview say that they have had this problem in the last 4 months.
Once you have done that, you go and execute the experiment. Of course, you can run more than one experiment at the same time if you wish.
First round of experiments.
Ok, so based on the assumptions I am making I decided to run a couple of experiments in parallel. If you have some comments I would love to hear them.What I chose to do was:
Problem Interview: Easy to do when your customers base is so approachable, and I know some people who fit the profiles. It yields the greatest learning for this part.
Landing Page stating UVP and sign up form to start collecting leads: Easy to do since I am using instapage.com (free) and lets you customise your landing easily as well as connect it to Google Analytics. You can read my other post if you want to know about other free tools I am using to build BARconWIFI.com.
Data Research of various kinds:
- Research on the Twitter channel to check on assumptions like: can I reach them, how can I deliver some of the value, are they even on twitter, etc.
- Estimate market size from the INE statistics and other sources.
- Find out if we can get the content to curate in sufficient amounts and quality.
This is all due to the fact that I am already paying for socialbro.com which allows me to easily run this experiments and research on twitter without much hassle and that I know how to use some other free tools to automate most of the process.
If that were not the case, I am not sure I would have gone this way.
Results.
In upcoming posts I will talk about some of the results for these experiments as well as some other details about executing the experiments. Until then, remember to share this article or leave a comment. :D
Posts
