PM Wargame: Lyft is considering a freemium model. Should they do it?
The board
Lyft introduced a bike product back in 2018. Since then, awareness and adoption have been slow. When consumers hear 'Lyft', they think 'car rides'; not bikes.
The senior leadership team is considering a freemium model.
As a product manager at Lyft, you're put in charge of figuring out whether this is a good idea or not.
In ‘PM Wargames’, I take on a hypothetical PM challenge and attempt to unpack the strategic & tactical approaches I would take to solve it. Unless sources are specifically cited, all information (including the challenge itself) is purely for illustrative purposes.
Actions
Business goals
There's a pretty big red flag on this board. It's one of those red flags that lie beneath the excitement of doing something new.
Why?
What's the actual goal of this initiative?
The first move I would make on this board is to walk over to the senior leadership team and get clarity on their goal. There are roughly three possible goals:
Raise awareness
Raise activation
Raise adoption
Goal 1: Raise awareness
The introduction tells me there's an awareness problem with our bike product. Customers know Lyft for its car rides, not its bikes.
A freemium programme could help raise awareness.
Goal 2: Increase activation
'Free' is a good reason to get people through the app and testing the product.
Beyond making them aware we even exist in this space, our goal could be to get users to go through several bike rides and reach our activation metric target.
Goal 3: Generate revenue
Finally, let's not kid ourselves: we want to make some cash. Freemium models are funnels after all, from free to paid.
I would try to get clarity on which of these three goals we are working towards with this initiative.
Research
While I leave the powerful people scratching their heads on the goal, I'll do some research of my own.
Customer research
First, I want to question the very core of this challenge. Are our customers really unaware of our bike product? Or are they just not interested? Have we asked them? When was the last time we asked?
Assuming we have data on this, I would dig through it to extract these insights.
If we don't have data, here's a customer survey I would immediately launch.
Segmentation questions
How many rides have you taken this past year?
Which product(s) do you use most (Lyft, XL, Transit, Bikes, etc.)
(I am assuming we can get basic demographic information from the customers' accounts).
Awareness questions
If you had to explain Lyft to a friend who's never heard of it, how would you describe our service?
Present a select few answers such as 'Car sharing service', 'Ride sharing business', 'An easy way to get around the city', etc. The goal is to see if customers truly associate Lyft with cars or transportation in general.
Did you know Lyft offer a Bikes & Scooters service?
Engagement questions
How many Bikes & Scooters rides have you taken in the past year?
What has your experience been on those rides (1-10)?
What, if anything, is preventing you from taking more Bikes & Scooters rides?
Present options such as 'Not enough coverage', 'Broken equipment', 'Too expensive', and so on. We want to find out if the problem with Bikes is technical or tactical.
Along with this customer research, I would get a small team of colleagues to pound the pavement with me. I want to ask similar questions to non-Lyft customers. I would go to universities, train stations, bus stops, etc. and ask the following:
Are you aware of Lyft? If so, can you tell me what it is?
Are you aware of Lyft's Bikes & Scooters product?
What has prevented you from taking a Lyft bike or scooter in the past year?
Usage
I am always sceptical of human-reported data. Although it's important to gather survey responses, I still want to check whether the usage data tells the same story our customers do.
I'll head over to our analytics team and ask for the following reports for the past year:
Total rides (across all products)
Total Bikes & Scooters rides
Total customers with at least one Bikes & Scooters ride / Total customers, expressed in % (customer penetration)
Bikes & Scooters rides per month (trend)
Bikes & Scooters ride ratings
Total Bikes & Scooters refund requests
Total Bikes & Scooters reports (defect, broken, etc.)
Finally, of course...
Revenue
I'll have an in-depth look at our revenue. How much is our Bikes & Scooters product generating right now?
Along with revenue, I want to segment our report in terms of:
Location. Which cities are our most/least active?
Day of the week. Which are our busiest/quietest days?
Time of day. Which are our busiest/quietest times?
Distance. How far do customers go on these products?
These segments will help us de-risk our initiative and craft a release strategy.
Game plan
Creating a freemium model is a risky strategy. While 'free' is great to attract new users, it's also, well, free.
There have been countless examples where 'free' broke the system. Equals went free and regretted it. CEO Bobby Pinero lives to tell the tale on Lenny's blog. Read.
The survivorship bias tends to draw us to success stories. It is reported Spotify & Slack both convert free users to paid at a staggering 30%+ (source). Sounds exciting!
In reality, I would expect:
For freemium to generally fail to generate much revenue. A good-to-great conversion rate from free to paid is ~3-8%.
For freemium to increase awareness but reduce revenue and increase costs. Free users still need support. Free users of physical products such as bikes may take less good care of the products, leading to more repair costs.
But those are assumptions and industry expectations. For this game plan, I will put these assumptions to the test with a clear focus on de-risking the initiative. I want to spot when things go wrong as quickly as possible so I can shut it down.
So here’s my game plan:
Given I haven't heard from my fake bosses about their goal with this initiative, I'll select 'awareness' as my initial goal.
It's a simple enough goal. I can run my test and compare usage & sentiment data again; and see which direction we are trending.
Select three test areas. I want to pick busy areas in large cities to get the maximum amount of data as quickly as possible.
Why not the whole city? To de-risk the trial. Offering free bike rides to all NYC residents is far too much risk.
Select two sets of off-peak hours during which rides will be free. I would expect those to be 10am-11:30am and 3pm-5pm.
Why off-peak times? Yes, to de-risk the initiative, but also to test the elasticity of the demand. If those are our least busy times, do free rides increase demand?
Model a timeframe for the test. I want to limit this test in time, perhaps three months before I review the performance.
Create a dashboard of our key metrics in these areas, at these off-peak times, with before/during/after monitoring.
Roll out the promotion to our existing customers via in-app notification.
Roll out a modest marketing campaign to attract new customers.
Premortem
I started this wargame with a red flag: we don't know why we're doing this.
I'll end it with the same warning. The main reason this test could fail (beyond freemium generally failing) is a lack of clarity on what we're hoping to gain from this initiative.
The game plan could (should) completely change depending on the goal. For example, if we're aiming to see an uptick in adoption and revenue, a 3-month trial won't be enough. To test our awareness goal, a 'modest marketing campaign' might not be enough.
My second concern will be around the hidden costs of running a freemium plan in the long run. Free customers still cost money to acquire and keep happy. They use and deteriorate equipment. They line up to speak to customer service.
To de-risk this, I would spend extra time with my friendly data analyst team to come up with some predictive set of metrics to keep an eye on. If things go south, I want to be able to pull the plug ASAP.
This is my game plan. What do you think of this approach? How would you play this PM wargame?