Find solution
How to find Stakeholder analysis tools?

How should a complicated business setup their business intelligence and analysis to make decisions best?

  • Many analytics packages and products to be appear geared to ecommerce companies where it is more straightforward to define 'lifetime value' since it's tied to purchase amounts and where most of the business truly is through a self-service platform. I am a product manager and want to do analysis to make decisions but there is a lot of grey area I'm having difficulty navigating. Although I'm the PM, the business stakeholders truly own the product and direction so I'm trying to arm myself with data to help guide and make an informed impact on the roadmap. The web app supports 3 customer types (people who make money via revenue sharing, people who spend money, and then people who make money and spend a percentage of it with the goal of making more money). There is also a decent percentage of users on both side who end up being hand-held (managed) throughout various steps of the process while others remain strictly self service. What should I track and what tools should I use to do so? I think I need to figure out how to answer questions like: -What is the lifetime value of customer type A, B, and C? -What is the impact of introducing feature x on each? -How much time does it take for customer type A to do x, then x, then ex, etc? -What is the value of the managed hand-holding versus remaining strictly self-serve? What cohort analysis should I be looking at? And most importantly, I'd like to define what the 1 metric for each customer type is that I should be looking at that can be used as an indicator of how valuable or invaluable a customer may be early in their life. For instance, Twitter determined that users who follow 10+ people in their first 24 hours end up being valuable/long-term users. How do I find that metric? My main goal is to be able to determine and quantify the impact of various things we're introducing (new features, marketing efforts, etc.) to figure out where to spend our resources in the future.

  • Answer:

    This is a pretty complex question and my primary advice is to bring in some analytics experts to help you identify the key metrics and how to measure them. Most analytics companies have consulting services they will provide along with their products and many (including ours) provide that service for free. However, in the meantime here are some guiding principles I recommend following whenever setting up analytics for your products: Quantity over Quantity. Pick 5 metrics that will be your KPIs (key performance indicators). The more metrics you have the more likely they are to become just numbers so you want to keep it simple. You also want to be able to focus on optimizing whatever KPIs you choose and that's hard to do with more than 5. Engineer for Success. You will optimize for any KPI you are tracking, whether you realize it or not. If you track revenue you will optimize for revenue, if you track margin you will optimize for margin - it's just human nature. Hence, make sure you choose whatever will make your business successful and monitor for that. The problem with "vanity metrics" (big numbers that feel good but don't impact your business) is that you will optimize for them at the cost of your business. Set Goals. Tracking KPIs is meaningless if you don't have a target. Maybe you want retention to be 5% higher in 3 months or you want revenue to be up 20% by the end of the year. Whatever it is set a goal, it will provide the focus you need. So how do you find that one perfect metric that works for your business? You experiment. Try a lot of different measures and see how well they represent the health of your business. It is very hard to find a single metric that encompasses everything about your business but some companies can do it. I find it much more common that you'll have 5 KPIs as described above. That being said, more practical advice for your specific questions; What should I track and what tools should I use to do so? Since you're not sure what will best serve your needs I would start with 3 tools, run them side-by-side and see what you prefer. For a web product like yours that likely means Google Analytics (http://www.google.com/analytics/), Mixpanel (https://mixpanel.com/) and KISS Metrics (https://www.kissmetrics.com/). Start with the basic integration and add more tracking as you realize you have more questions they don't answer. What cohort analysis should I be looking at? The two most popular cohort analyses that are done are cohorts by acquisition date (when did they first start using your service) and cohorts by spending level (how much money do they spend with you). Sometimes the latter is broken up by engagement if it's a free service.

Sean Byrnes at Quora Visit the source

Was this solution helpful to you?

Other answers

Great answer by . I'd like to add to his first paragraph that from your post alone, I feel you are a bit overwhelmed by the complexity of the problems at hand and that is a bad place to be at. More important than advice, you want some consulting. Someone that digs in your business and identifies what and how to measure. I won't get into the platforms since I'm in the mobile games space which is, in my opinion and experience, tremendously different from web apps. I would advise to pick one that has AB testing included. I'll do my best to help you out with your questions but please consider getting someone to work closely with you regardless of what you think of my help: -What is the lifetime value of customer type A, B, and C? The analytics platform must support user segmentation based on some parameter passed. Divide the sum of revenue by the number of users and segment by the customer type parameter and you'll have the average revenue per user of each segment. Going from that to LTV is a matter of model choice. To get the revenue you need a purchase event, to get the number of users you need a session event of some sorts. Both of them with unique user IDs. -What is the impact of introducing feature x on each? This is achieved by randomised trial aka AB tests. You should be able to know this kind of stuff with the session and revenue events on top of a platform that supports AB testing. -How much time does it take for customer type A to do x, then x, then ex, etc? This is a funnel. You can achieve this in a number of different ways but the basic setup is to send an event every time a user changes type (including type A! this is important) or to send the user type as a parameter in the session event. I'd prefer the first since it is more granular but second is easier to implement usually. -What is the value of the managed hand-holding versus remaining strictly self-serve? I don't understand this question. -What cohort analysis should I be looking at? The single most important one is retention. In web analytics terms, it is the opposite that it's measured and it's called churn. This is what Sean refers to as by acquisition date. And most importantly, I'd like to define what the 1 metric for each customer type is that I should be looking at that can be used as an indicator of how valuable or invaluable a customer may be early in their life. For instance, Twitter determined that users who follow 10+ people in their first 24 hours end up being valuable/long-term users. How do I find that metric? This is inference/predictive analytics stuff. Don't do this by yourself. You need to be able to have a full user state with very well defined and implemented events that define the user behaviour. After having this, you must model data in a way that you get the desired result, either an interpretable model from which you can define future actions or a machine learning algorithm from which you can extract the highest possible prediction accuracy at the expense of fully understanding what is behind it. Whatever it is, it is a complex task that goes from defining the possible user attributes, how to get them, how to get the dataset, model and validate and then apply. It's a job for a data analyst, or machine learning practitioner or a data scientist and it is very hard.

Ricardo Vladimiro

Related Q & A:

Just Added Q & A:

Find solution

For every problem there is a solution! Proved by Solucija.

  • Got an issue and looking for advice?

  • Ask Solucija to search every corner of the Web for help.

  • Get workable solutions and helpful tips in a moment.

Just ask Solucija about an issue you face and immediately get a list of ready solutions, answers and tips from other Internet users. We always provide the most suitable and complete answer to your question at the top, along with a few good alternatives below.