Lean Mastery Collection Page 6
There are a few additional types of pivots as well. A Pivot that zooms in is one that takes a signal successful feature of a failed prototype and turns it into its entirely own product. A zoom out pivot, on the other hand, is when a failed prototype is useful enough to become a feature on something larger and more complicated.
The customer segment pivot occurs when the prototype proves solid, but the target audience proves to be different than anticipated. A customer need pivot occurs when it becomes clear that a more pressing problem for the customer exists, so a new product needs to be created to handle it.
A platform pivot occurs when a single application becomes so successful that it spawns an entire related ecosystem. A business architecture pivot occurs when a business switches from having low volume and high margins to high volume and low margins. A value capture pivot is one of the most extreme as it involves restructuring the entire business to generate value in a new way. The engine of growth pivot occurs when the profit structure of the startup changes to keep pace with demand.
Small batches: When given the option to fill a large number of envelopes with newsletters before sending them out, the common approach is to do each step in batches, fold the newsletters, place them in the envelopes, etc. However, this is actually less efficient than doing each piece by itself first, thanks to a concept known as single piece flow, a tenant of Lean manufacturing. In this instance, individual performance is not as important as the overall performance of the system. Time is said to be wasted between each step because things need to be reorganized. If the entire process is looking at a single batch, then efficiency is improved.
Yet another benefit to smaller batches is that it is easier to spot an error in the midst of them. For example, if an error was found in the way the envelopes were folded once all the newsletters had been folded, then that entire step would need to be repeated, adding even more time to the process. On the contrary, a small batch approach would determine this error the first time all the steps were completed.
Andon cord: The Andon Cord was used by Toyota to allow any employee on the production line to halt the entire system if a defect was discovered at any point. While this is a lot of power to give to every team member on the floor, it makes sense as the longer a defect continues through the process, the more difficult and costlier it will eventually take to remove. As such, spotting and calling attention to the problem as quickly as possible is the more efficient choice, even if it means stopping the entire production line until the issue is fixed.
Continuous deployment: Continuous deployment is one of the most difficult Lean Startup processes for many companies to deal with as it means constantly updating live production systems each and every day until they reach an ideal state. The essential lesson is not that everyone should be shipping fifty times per day, but that by reducing batch size you can make it through the entire build, measure, and learn cycle more quickly than your competition can. The ability to learn directly from customers is essential in this scenario as it is one of the primary competitive advantages that startups possess.
Kanban: This is another part of the process that is taken directly from Lean manufacturing. Kanban has four different states. The first of which is the backlog which includes the items that are ready to be worked on but have not yet been actively started on. Next is in progress, which is all of the items that were currently under development. From there, things move to build after development has finished and all the major work has been done so that it is essentially ready for the customer. Finally, the item is validated by a positive review from the customer.
A good rule of thumb is that each of the four stages, also known as buckets, should contain more than three different projects at a time. If a project has been built, for example, it cannot then move into the validation stage until there is room for it. Likewise, work cannot start on items in the backlog until the progress bucket has been cleaned out enough to free up the space. One outcome that many Lean Startups don’t anticipate is that this method also makes it easier for teams to measure their productivity based on the validated learning from the customer as opposed to the number of new features being produced.
Five whys: Many technical issues still have a root at a human cause at some point in the process. The five whys technique makes it possible to get close to that root cause from the beginning. It is a deceptively simple plan, but one that is extremely powerful when used by the right hands. The Lean Startup system posits that most problems that are discovered tend to be the result of a lack of personal training, which on the surface can either look like a simple technical issue or even one person’s mistake.
For example, with a software company, they may see a negative response from their customers regarding their most recent update. Looking more closely at the issue, it was discovered that this was due to the fact that the update accidentally disabled a popular feature. Looking closer still, this was discovered to be due to a faulty service which failed because a subsystem was used incorrectly due to an engineer that wasn’t trained correctly. Looking closer still, you will find that this is due to a fact that a specific manager doesn’t believe in giving new engineers the full breadth of training they need because his team is overworked and everybody is needed in one capacity or another.
This type of technique can be especially useful for startups as it gives them the opportunity to determine the true optimum speed needed to make quality improvements. You could invest a huge amount in training, for example, but that doesn’t mean this is always going to be the right choice at the given stage of development. However, by looking closely at the root causes of the problems in question, you can more easily determine where there are core areas that require immediate attention as opposed to solely focusing on surface issues.
Another related issue is connected to the fact that many team members are likely prone to overreacting to things at the moment, which is why the 5 Whys are useful when it comes to taking a closer look at what’s really happening. There can be a tendency to use the Five Whys to point blame, at first, but the real goal of the Five Whys is to find any chronic problems caused by bad process, not bad people. This is also important to ensure that everyone is in the room together when the analysis takes place because it involves all of the people impacted by the issue, including both customer service and management. If blame has to be taken, it is important that management falls on the sword for not having a team-wide system in place to prevent the issue in the first place.
When it comes to getting started with the Five Whys, the first thing that should be focused on is instilling a feeling of trust and empowerment in the team as a whole. This means being tolerant of all mistakes the first time they happen, while at the same time making it clear that the same mistake should not happen twice. Next, it is important to focus on the system level as most mistakes are made due to a flaw in the system which means it is important to put the focus on this level when it comes to solving problems.
From there, it is important to face the truth, no matter how pleasant or unpleasant it might be. This method may bring up some unpleasantness about the company as a whole but the goal is to fix these issues, after all, and you can’t fix what hasn’t been brought to light. This is why it is easy to turn it into the Five Blames if you aren’t careful which is why the blame should flow up in this instance. Start small and be specific. You want to get the process embedded, so start with small issues with small solutions. Focus on running the process regularly and involving as many people as you can.
Finally, it is important to designate one person on the team as the Five Whys Master. This person will be the one who is primarily in charge of seeing that change actually comes to the team. This, in turn, means they will need a fair amount of authority in order to ensure things get finished. This person will then be the one accountable for any related follow-up, determining if the system is ultimately paying off, or if it is better to cut your losses now and move on. While it can ultimately be a great way to
create a more adaptive startup, it can also be harder to get into the groove of than it first appears, so it is important to look at it as a long-term investment rather than something that will be completed in the short-term.
Chapter 2: Create a Useful Lean Startup Experiment
Qualitative or Quantitative: While many people assume that their startup experiment needs to be either quantitative or qualitative, the fact of the matter is that one is not inherently superior to the other. Instead, it is better to think of the two as if one was a hammer and the other was a screwdriver. While a hammer is better at putting nails in wood, that doesn’t mean it is inherently superior on all fronts. Any tool can be used for good or evil, which is why it is important to focus more on validating the right metrics than it is to worry about which of these two processes is superior. In fact, using qualitative research and then validating it with quantitative research is likely going to do the most good anyway.
Generative or Evaluative: A generative research technique is one that doesn’t start with a hypothesis per se but can still result in a wide variety of different ideas. Things like Customer Discovery Interviews fall under this type of technique. Evaluative, on the other hand, is all about testing a very specific hypothesis in order to determine a very specific result. The popular smoke test falls under this type of testing. It is perhaps this distinction, more than any other, that explains why some people end up with poor results from their experiments.
For example, a smoke test could be run to test the hypothesis that some percentage of the market will be interested in shoes that are compostable. To test this hypothesis, you would then put up a fake coming soon landing page explaining that compostable shoes are totally going to be a thing and see who signs up for the newsletter. After the work was done and the results were in, it turns out that there was about a 1 percent conversion rate when it comes to the shoes. The good news is that the hypothesis was confirmed, the bad news is that it wasn’t particularly useful.
What’s more, the results are unclear because it still isn’t clear if the interest isn’t there, if the advertising was poor, or if there is a third variable that you aren’t yet aware of. This can be broadly defined as the difference between people not being interested in the value proposition and people not understanding it. The truth of the matter is that there are hundreds of reasons out there why someone might get a false negative result from a given test, just as there are a number of reasons why a false positive might be generated.
To get started, you will need to determine if the hypothesis is flawed or simply vague and, in this case, it is both. Some people are too vague when it comes to a target audience, some are a poor qualifier. As such, first, you would need to focus on a more specific demographic, and second, you would need to do research to determine how big the audience for compostable shoes would ultimately be. Only once the hypothesis is falsifiable and specific can it benefit from an evaluative experiment like the smoke test. If you can’t clear up your hypothesis then you will want to start with Generative Research and work back from there.
Market or product: When it comes to the distinction between methods and tools, the biggest is perhaps the distinction between Product and Market. Some methods are useful when it comes to helping startups learn about their customers, their problems, and their best lines of communication. As an example, startups can listen to their potential customers to make it easier for them to understand their specific situations and what their day to day problems are like.
Other methods make it possible to learn about the product or a potential solution that will help to solve a specific problem. One good place to start is with a set of wireframes as a means of determining if the interface is as usable as it seems at face value. Unfortunately, this still won’t make it clear if anyone is going to buy anything in the first place.
As these methods don’t typically overlap all that well, it is important to choose one and stick with it throughout its cycle. If you combine evaluative research and generative research with Product and Market, you will end up with four different means of determining the best path forward.
Generative Market research asks questions like:
Who is our customer?
What are their pains?
What job needs to be done?
Is our customer segment too broad?
How do we find them?
If you can’t answer these questions clearly and easily, then your startup is in what is known as the Customer Discovery phase. During this phase, it is important to get to the basis of the problem prior to testing out any potential solutions to ensure that you are actually solving the right problem in the end. If you don’t have a clear hypothesis to start, then you will need to generate ideas.
To do so, you may want to talk to customers to see what is bothering them or you could use a data mining approach to determine the problem, assuming you have access to enough data. You may even want to use a survey with open-ended questions if you are really fishing for ideas. Some of these methods will be qualitative and some will be quantitative, but this distinction is ultimately irrelevant in the long run. Data mining is a quantitative approach, but it helps identify problems, most famously the existence of food deserts which would have been difficult to determine in virtually any other way.
Generative Market Research Methods include:
Surveys
Focus groups
Data mining
Contextual inquiry / ethnography
Customer Discovery Interviews
Evaluative Market experiment questions include things like:
How much will they pay?
How do we convince them to buy?
How much will it cost to sell?
Can we use scale marketing?
In order to properly evaluate a specific hypothesis, you may want to start with a landing page to determine if there is likely to be a demand. You may want to put together a basic sales pitch if you are working on a B2B enterprise type product. You could even go so far as to run a conjoint analysis as a means of further understanding the relative positioning of a few value propositions.
Evaluative market experiments that are useful if you have a clear hypothesis include:
High bar
Fake door
Event
Pocket test
Flyers
Pre-sales
Sales pitch
Landing page
Video
Smoke tests
Surveys
Data mining/market research
Conjoint Analysis
Comprehension – link to the tool
5-second tests
While this sort of research can provide lots of interesting data, it is important to keep in mind that much of it still has the potential to be wrong as signing up for a landing page is very different than actually putting money down on a product. In any situation where the customer doesn’t have to commit anything more than an email address, then they don’t signify an actual customer demand.
It is important to keep in mind that the value proposition and the product are not the same things. The value proposition is the benefit that your product will deliver to your target audience. As such, you cannot have a validated value proposition if you don’t have a validated customer segment.
Chapter 3: Growing a Startup
When a startup is composed of only a few people, the small team that started the company, it’s easy to manage everyone and everything. You’ve got your first few clients, and they’re happy with your work, paying all their bills on time and referring your services to other potential clients. But as your startup grows—with more staff, more clients, and more money to keep track of—it can be a challenge to manage all these aspects efficiently.
But there are ways to make this process easier so that you don’t lose too much time or money. It’s all about ensuring you use collaboration, effective lead generation, and strict budgeting. Here’s how:
Collaboration
Cage is a new platform that enables contextual collaboration. Through Cage, teams have the ability to gather feedback in real time, assign tasks, edit images, and distribute media files, all on one platform. By facilitating the entirety of a project, from the initial brainstorm to a final review before a video or platform is published, Cage ensures that everyone involved in the projects has full insight into updates and strategic pivots.
Regardless of the medium, every project takes on a life of its own. More often than not, facets shift over time, and these changes and discussions are often implemented across several platforms, which often lead to confusion and oversight. Cage helps teams avoid this pitfall and, as a result, empowers them to collaborate more effectively and efficiently.