Lean Mastery Collection

by Jeffrey Ries

  Segmentation

  The first test is segmentation. This process involves comparing a set of data from a demographic bucket. You can divide up the demographics in any manner that you choose such as gender, lifestyle, age, or where they live. You can use this information to find out where people are purchasing a particular product; if there are different buying behaviors between female and male customers; and if your target audience seems to be in a certain age group or not.

  The reason that you want to build up a user segment is to make it easier for the data to be actionable. Analytics can teach you a ton about the people who purchase from you, but there is often a lot of information there, and it can be hard to draw good conclusions from this information. After all, while this information from the past can be useful, it isn’t going to be the best to tell you how to improve either retention or conversion rate.

  This is where the process of segmentation is going to come into play. When you learn how to filter out the audience, you will then be better able to create a plan to make new products that serve them the most. Analytics can give you the information that you need, but segmentation can help you to act.

  For example, you may have a conversion rate that seems average or good, but it could be from a combination of one group that converts really high and consistently so, and then another group that seems to never convert at all. You could be wasting a lot of money on that second group where you are hardly getting anybody to convert at all. Segmentation can be used to help you understand what things you are doing the right way when engaging the first group, and can give you a plan on how you can improve to work on that second group.

  With segmentation, you don’t want to only look at the data to learn some more about your users, but you also want to come up with data that you can act upon. Segmentation can help you with this. You will be able to divide up the people in your customer base and learn how to advertise to them better than ever before.

  Remember that not all customers are going to be the same. There are some of your customers who may purchase something once, and they aren’t regular customers. While it is still good to reach out to them, you want to learn who your regular audience is, what they respond to, and what keeps them coming back. This is going to ensure that you keep them coming back and earn as much profit as possible.

  So, how do you create a segment of your customers? There are many different options that you can use when creating a segmentation. But let’s look at the process that you can use to create a segmentation for your Lean Analytics project. The steps you need to use include:

  Define the purpose of your segmentation: You should first figure out how you want to use your segmentation. Do you want to use it to get more customers? Do you want to use it to manage a portfolio for your current customers? Do you want to reduce waste, become more efficient, or something else? Defining your purpose can make it easier to know how you should segment out your customers.

  Identify the variables are the most critical: These are going to help influence the purpose of your segmentation. Make sure that you list them out in order of their importance, and you can use options like a Decision tree or Clustering to help. For example, if you want to do a segmentation of products to find out which ones are the most profitable, you would have parameters that are revenue and cost.

  Once you have your variables, you will need to identify the threshold and granularity of creating these segments. These should have about two to three levels with each variable identified. But sometimes you will need to adapt based on the complexity of the problem you are trying to solve.

  Assign customers to each of the cells. You can then see if there is a fair distribution for them. If you don’t see this, you can look for the reasons why, or you can tweak the thresholds to make it work. You can perform these steps again until you get a distribution that is fair.

  Include this new segmentation in the analysis and then take some time to look it over at the segment level.

  Cohort Analysis

  The Cohort Analysis is a test involves comparing sets of data using a time bucket. In this test, there will be differences in behavior between customers who arrived at the free trial stage of your process, versus those who showed up at the initial launch, and then those who are in the full payment stage.

  Each of these is significant because it helps you to figure out which customers are likely to come back and be full-fledged customers when in the future. Those that show up in the initial stages when the product is free are often not the customers you are going to see when sales actually start. They may have just wanted to try it out and didn’t really have an investment in the product.

  Those that are in the later two stages can be customers who are better for you to work with. They will be the most interested in the product because they invested some money to get it. You really want to study these using the cohort analysis to figure out who your real customer base is and how they behave so that you can better market to them later on.

  A/B Tests

  A/B testing is a process where you examine an attribute between two choices. This could be something like an image, slogan, or color so that you can figure out which option is the most effective choice.

  Let’s say that you had two products that you are comparing and you want to find out which ones customers liked the best. Did they choose one product over the other and why? Did they respond better to the choice that was in green or the one in blue?

  For this test to really work, you must assume that everything else is going to stay the same. So, it would have to be the same product, but there is one variable that is different between them. You could put up a website, for example, and have a red background on one version and a yellow background on another. Then you could use A/B testing to figure out which one the customer responded to the best out of those choices.

  In addition, you can also work on multivariate analysis. This is pretty much the same thing, but instead of going through and testing out one attribute, you will go through and compare several changes against another group of changes to see which is the most effective. This one will require there to be a few changes in the second product compared to the first to be the most effective.

  There are several keys that you need to have in place when you are ready to do an A/B test. These include:

  Know the reason that you are running this A/B test.

  The item that you are testing needs to be noticeable to the audience. If you make a minor change that no one is going to notice, then your results are not going to be that reliable.

  Stick with testing just one variable at a time. If you go through and do multivariate testing, or test more than one thing at a time, you will run into trouble. You may not know for sure which variable is causing the changes you see.

  Your test needs to end up being statistically significant. This means that it must have a sample size that is big enough to test and know that the results are valid within a certain margin of error.

  Let’s take a look at an example of how to do this. We are going to use this test on a website that you are trying to improve. There are two main ways that you can do this including:

  You will test the pages on separate pages.

  You will use JavaScript to conduct the test inside the page, so you don’t need two different URLs to do it.

  The first option is going to mean that you will need to have two different URLs for the pages that you are testing out. You can make them similar names, but make sure that there is some way that you can keep track of them and not get the two confused.

  With the second option, you will need to have some experience working with JavaScript. You can then place some of this code on the website so that it can dynamically serve one option or the other.

  The method that you choose is often going to depend on the one that you like the most and which tools you want to use. Both of these will give you some valid results, but you will find that implementing each of them takes a different amount of t
ime to set it up.

  Chapter 7: Step 1 of the Lean Analytical Process: Understanding Your Project Type

  Now that we have taken a look at some of the different part of Lean Analytics, it is time to take a closer look at how the process works. These can help you to get started with the Lean Analytics stage for your business and ensure that you are getting the most out of Lean.

  The first step that we are going to look at is understanding your business or your project type. How are you supposed to pick out the right metrics if you have no idea what kind of business or project type you are working on? You must really understand the project at hand so that you can choose a fantastic metric that can show you results.

  There are six general business types that you can fit into, and they all will have metrics that are going to work best or matter the most, for each one. If you see that your business or project is on this list, your job will be simple. You just need to focus your attention towards understanding the priorities of what needs to be measured. This can include in-depth external research.

  However, if you have a business that is not on this list, this doesn’t mean you are out of luck and can’t do anything. You can just use some of the information that is in this chapter as an example and build up your own understanding and metrics from this chapter.

  E-commerce

  The first type of business is going to be e-commerce. These are growing like crazy right now as many customers are looking for the things they want to buy online more and more. And many companies find that they can make large profits by offering their products and services online to these customers. An e-commerce business is going to be any that has their customers buy from a web-based store. This could include businesses such as Expedia.com, Walmart.com, and more.

  The strategy for this type of business is that you need to understand the customer relationship that you want. This means that you are going to focus either on new customer acquisition or customer loyalty? You have to decide between these two because this is going to help with all other decisions that you make with this type of business.

  There are many metrics that you can choose to go with in an e-commerce business. Some of the typical ones that other companies have chosen in this industry include:

  Inventory availability

  Shipping time

  Mailing list and how effective it is

  Virality

  Search effectiveness

  Shopping cart abandonment

  Revenue that you make on each customer

  The amount you spend to get new customers

  Shopping cart size

  Repeat purchase

  Conversion rate

  The best metrics

  Of course, there are several metrics that will work the best and will provide you with the best return on investment, when working with an e-commerce site. The best metrics to use here include:

  Conversion rate: This is the percent of all visitors to your site who also purchase something. The average conversion when it comes to online retail is 2%. There are some that can do better though. For example, Tickets.com is over 11%, and Amazon.com is at almost 10%.

  Shopping cart abandonment: It is typical that 65% of the shoppers to your website are going to abandon their carts. Many of these are because of the high costs of shipping, and others are from the high price of all the items in their cart. You should definitely take some time to analyze any shopping cart abandonment that is happening in your business so that you can learn why you are losing these customers.

  Search effectiveness: The majority of your buyers are going to have to search to find what they need. If you make your search more effective, it can help your customers find what they want, rather than having them leave in frustration. Remember that about 79% of your total shoppers will use the search engine for half of the goods they want.

  Software as a service

  These types of companies are going to sell software in downloadable form or as a subscription. This can be things such as Skype, Evernote, Basecamp, Adobe, and more. They are not selling a physical product to someone, but these software programs are still pretty important for most people to get work done or to do other things on their computers.

  The strategy with this one is that most software is going to consist of products that are on a subscription which means that retaining the customers is important. Your success is going to really depend on building up a loyal base of customers faster than those customers disappear.

  There are some metrics that you can use to make this happen. Some of the most common metrics that are used with this type of business model include:

  Reliability and uptime

  Upselling

  Virality

  Customer churn

  Customer lifetime value

  Cost of getting new customers

  The amount of profit you make per customer

  User conversion

  User stickiness

  User enrollment

  User attention

  The best metrics

  Just like last time, you are able to use any of the metrics that are above, but there are some that could be the best for helping you reach your overall goals. Some of the best metrics to use with a software company includes:

  Paid vs. free enrollment: You will find that your enrollment rate is going to change depending on whether or not you asked for credit card information in the free stage or not. The former is going to get an average signup rate of 2 percent, and then 50 percent often end up buying. When you do not ask, the average may increase to ten percent, but only 25 percent purchase the product.

  Growing revenues and upselling: Some of the best software providers are able to get 2 percent of their paying subscribers to increase what they pay each month. Being able to grow your customer revenue by 20 percent in a year can be achieved if you work towards it.

  Churn or attrition rate: This is the percent of your customers who are leaving. Going across the industry, the top companies usually have an attrition rate between 1.5 and 3 percent each month. If you have a percentage that is higher, then you need to find ways to make the customers stay.

  Mobile app companies

  These are companies that are going to provide apps to be used on mobile devices like Android and iPhone. Some of the companies that can fall under this category would be ones like WhatsApp and Instagram.

  The strategy that you want to go with here is to find the right target audience. There are a lot of ways for your app to make money, but you will find that the majority of your revenue is going to come from a smaller group of customers, rather than from the population as a whole. You should focus your analysis as well as the metrics you use to help you better understand those customers.

  There are many metrics that you are able to use as an app company. Some of the most common options include:

  Customer lifetime value

  Churn rate

  Ratings click-through

  Virality

  The revenue you make from each paying user

  The revenue you make for each user

  Percentage of users who end up paying

  How much it costs to get the customers

  Launch rate

  Downloads

  Best metrics

  Of course, there are many metrics that you can choose to look at when it comes to being an app company, but a few of them are going to provide you with the most information and can help your business to really grow. Some of the best metrics you can use include:

  Downloads and the app launches: The number of people who download the product and then activate it will fit in here. It is known that quite a few people who decide to download an app will then never activate it or use it at all, especially if the app is free.

  The cost to get new customers: You can follow a general rule to have a budget of 75 cents per user in your marketing initiatives to help attract new customers. You should always make sure that the cost to get new customers is lower than what you w
ill earn on them. So, if you will only earn 50 cents on a customer, then you shouldn’t spend 75 cents on each one.

  The average revenue you earn per customer: This is often going to be determined through the business model. For example, Freemium apps, or apps that you receive revenue from engagement in the app, will often have a higher revenue per user compared to those that are premium apps.

  Media site companies

  If you are in this industry, you have a website that is going to provide some information, such as articles, in return for earning advertising or any other type of revenue. These would include most blogs and other sites like CNN.com, CNET, and more.

  Media sites need to really understand the source of their revenue. It is not coming directly from their readers or the people who use their “product,” but it is coming from advertisers who are trying to reach those readers. So, if you are a media site company, you would get revenue from affiliates, click-based advertising, display advertising, and sponsorship. You would want to design your key metrics to work for this.

  Some of the different metrics that you can choose to work with for a media site company include: