By observing other parts of the Share Now service offering, we noticed that there were ample use cases for providing additional extras for longer trips. With the increased traffic on our hourly and daily rates for spontaneous rentals, we decided to go after this potential revenue.
2020 - 3 Month
Responsibilities
Stakeholder management
Tools
Stakeholder
◇ The Goals
◇ Starting Point
Aside from the spontaneous car-sharing offer where individuals can pick up a car anywhere on the street and drive, Share Now also offers a service for more planned trips. In this pre-booked environment, experiments around upselling trips with e.g. better insurance, additional drivers, etc. have proven wildly successful.
We were able to increase the revenue generated via pre-booked trips by roughly 8% and wanted to leverage similar effects for the spontaneous service offering.
◇ Understanding the Status Quo
2 main types of use cases with vastly different expectations towards our service:
The commuter
Minute Rate - 12 hours
- About 80% of Share Now users, especially on trips between office and home
- Jobs to be done: Get from A → B ASAP
- Low sense of risk due to short distance in a familiar area
Priority on Speed
The traveler
1 - 30 days
- Growing user segment Learn how a redesign increased this use case
- Feeling towards ownership over Share Now car for the duration of the trip
- Longer trips in less familiar territory have a higher need for upgrades
Priority on Comfort + Security
Provide appropriate extras
◇ Exploring options
To find the best possible solution for presenting the appropriate extras to a user while considering their use case and jobs to be done, we started with creating wireframes.
To reduce friction in the delivery, later on, we included engineering in the creation of these concepts.
With this feature only being available in certain markets, we needed to find a way to weave it into the existing screen’s layout.
The existing screen’s layout without any upselling options on the right provides the base canvas for this feature →
Dedicated Step
User will be sent to a designated page for all extras regardless of what rate they select. The available options would be according to the selected rate.
Extras Pills
No designated step in the flow. User’s can opt into selecting extras available on a given rate by tapping the respective pill on the bottom sheet.
Clustered Rates
Available options are located on the top of the available rates. Rates are clustered into hourly and daily rates to allow for an appropriate selection of options on the subpage
◇ Testing
Round 1 (peer review)
We floated the different ideas among engineers to learn about technical challenges. The response was pretty positive with the exception that clustering the rates into hourly and daily would be a bit more involved than pursuing concepts with the current presentation of the rates.
Therefore we created UI iterations of the Extras Pills vs the Dedicated Step concepts and showed them to our fellow designers. The goal was to uncover pain points in the respective flows as well as learn more about the discoverability of the new feature considering the POV of heavy users.
Extras Pills
→ Play with Prototype
Dedicated Step
→ Play with Prototype
Learnings
A: Extras Pills
The extras pills worked rather well. Once discovered, the participants understood the pills quickly and adjusted their trip options according to the test scenario given.
The connection between different rates and their corresponding extras was not easily understood
A downside from this concept was the poor scalability. Already with only 3 available extras, the affordance and ability to maintain an overview was challenging.
The additional costs for the chosen extras are not transparent enough.
B: Dedicated Step
There were some issues with this version.
The insurance option under minute rates are as easily discovered as with the pill concept.
When selecting an hourly or daily rate, the options (including insurance options) are hidden in the forced extra step. This extra step’s affordance is limited given that it only manifests in a change in CTA copy.
The additional costs for the chosen extras are not transparent enough.
Round 2 (in-house user testing)
Based on the feedback from our colleagues, we decided to chance both versions slightly before going into testing with non-designer coworkers.
A: Extras Pills
To strengthen the connection between specific rates and available extras, we connected the two elements visually.
With the rate cards now expanding on selection to show available extras, we assumed the affordance to increase as well.
Having learned about the intransparency related to cost, we augmented the bottom sheet with the sum of the extras.
The extras sum was intended to work as an entry point into the selection of extras
→ Prototype
B: Dedicated Step
The main change with this version was the visual unification of the bottom sheet.
One main concern with this change was a potential decrease in affordance for extra insurance options available for minute rate rentals (main business case). In this scenario, the entry point to upgrade the insurance was indicated with nothing but a small icon under the price 🙈
→ Prototype
Scope and test setup
Learnings
Good stuff
💊 Generally preference towards the version with pills
📈 Pills raise curiosity and after a first interaction are immediately clear in their meaning
🎯 Good task success with the pill version
💨 Appreciation of unchanged flow for minute rates (extras selection optional)
Bad stuff
❓ The extra entry point in the bottom sheet is confusing
❓ Pills have little affordance as to what’s behind them initially (fast learning curve thought)
💸 The pricing structure is not clear enough
🔎 The dedicated step makes it hard to experiment with the cost
◇ Hybrid version
It had manifested, that a forced extra step to upsell extras was considered disruptive for minute-based rentals. On the other hand, we wanted to make sure that users selecting hourly or daily rates were aware of the options available to them. Therefore we discussed a more flexible logic that ensured at least one extra touchpoint for users renting a car for longer periods.
Based on the previous rounds of testing, we identified 3 main target areas to improve on before releasing the feature into the wild.
→ Final Prototype
◇ Success?
Maybe! We set a very ambitious target for this feature ... maybe without thinking it through.
Why was it so ambitious in hindsight?
We know that the majority of trips done spontaneously are still minute-based rentals from commuters. This user group is unlikely to purchase any extras.
Looking at the trip data, users rarely rent a car longer than 1 day spontaneously. Longer trips are mostly planned in advance and therefore pre-booked.
With the assumption being that the shorter the trip, the less need for additional extras, the revenue stream to be expected from spontaneous trips is limited.
With 10% of trips opting for mainly upgraded insurance, we were however able to generate ~2,4% of additional revenue from this feature. This is well below target, but given the overall revenue size of our business not a negligible amount nonetheless.
◇ Next steps
That is a tricky question. With the learnings from working on this feature, it is questionable how much of a use case there is for longer trips started spontaneously. Hence, how much revenue can be generated from extras sold in this environment?
It is my position, that to increase the revenue stream of spontaneous extras, the use case for longer spontaneous trips needs to be established.
If we could prove this hypothesis right and increase the utilization of our spontaneous longer-term offers, we could continue with improving the upselling by offering tailored extras to users’ use cases or find even better ways of increasing our revenue while improving the usability of our service overall.
◇ Key Take-Aways
The initial assumption about this feature’s business impact was based on a rather broad observation. We had increased the revenue generated by non-minute rate-based rentals quite significantly and thought it was only natural for the upselling options to perform similarly in the spontaneous context than in pre-booking. However, had we looked a bit deeper into the numbers initially, we could have learned early on that the average trip duration is much shorter in spontaneous rentals and therefore the probability of significantly upselling rides was limited.
I still would consider the release of this feature a success as a 2% revenue increase is nothing to be sneezed at. However, with the expectation for improvement being quite inflated, not reaching the astronomically high set target leaves a bad taste in everyone’s mouth.
This project proves once again that dwelling in the problem space and diving deep into the discovery before starting to work on solutions usually yields better results after all!