The first basic rule of a Lean Startup is
IF THEY COME YOU SHOULD BUILD IT...
Not if you build it they will come.
Don't make what you can sell, make what you can sell.
Silicon Valley - they hate building something that nobody wants...waste of your time and energy.
The core of the Lean Startup process is
- Have and idea and build something
- Measure what that does in the market
- and then use that to learn
Your FIRST product is not a product to be successful as a product.
Its a tool to learn what you should have built in the first place.
We all delude ourselves into thinking our idea is awesome.
The risk is not CAN YOU BUILD IT?
The risk is WILL ANYBODY CARE?
Things fall apart in the measure part because
- You didn't measure things
- You are so delusional about your idea that you think you don't need to measure things
- Measurements are so sad, you sweep them under the rug and hope they get better later.
No measure means you didn't learn anything...no data, no learning, no next product.
Most successful products started of building something else.
Eg: Flickr was going to be multi-player online game - but people started uploading pic > sold to yahoo.
Flickr team went back to build that video game thing - ending up building Slack to coordinate that distributed team.
Products have no idea what they are going to be, and Analytics is going to help you figure that out.
Analytics - is the measurement of movement towards your business goals
In a startup the reason you use analytics, is to find the product market fit.
That magical unicorn - I have figured out what to SELL to people in a certain market I can reach before the money runs out.
Analytics is about tracking the right metrics.
A good metric has 4 fundamental attributes.
- A good metric is understandable
- If you are explaining the metric to people...they are not talking about the business they are talking about what the metric means.
- Should be something like a golf score - I was supposed to get it in Hole in 5, I got there in 7 - easy to understand.
- A metric must be comparative
- How did we do against last week, last month, against competition.
- A metric must be a ratio or a rate.
- A METRIC HAS TO BE BEHAVIOR CHANGING (SINGLE MOST IMPORTANT THING)
- If a metric won’t change how you behave, it is a bad metric.
Lots of people create numbers because it makes them feel like they are doing something.
eg: If its goes over 5%, we are going to do this.
E-Commerce example - Super important, almost nobody measures it.
Measure the number of customers who come back and purchase from you a second time in 90 days. (very simple to measure)
Say <15% buy from you a 2nd time in 90 days ----> It means, you are an acquisition focused e-commerce provider (not keeping them around - ok if you are selling wedding rings)
You are going to get that person once, charge them as much as you can - insurance, etc.
Say >30% buy from you a 2nd time in 90 days → Maybe you are a pizza business, you are in loyalty business. One individual pizza is not that important, so you can offer free if you screw up.
Qualitative Vs Quantitative Metric
Qualitative - that pizza is good, it is soft > useful in discovering the things your customers need.
Quantitative - I finished it in 7 bites > hard numbers, facts
Exploratory Vs Reporting Metric
Exploratory - I don't know what I am looking for, just digging around. Looking at data to look for ideas.
Reporting - Static data
Cohorts & Segments
Cohorts - A group of people that arrived at a particular amount of time. Looking at people who experience the March product, or April product.
Segments - multi variant analysis (A/B)...see what combination of buttons, colors etc work best.
People don't realize how important Cohort analysis is because if you use averages, you are dragging down your numbers with past group who experienced an inferior product compared to your latest and greatest.