At the core product marketing at Uber, we focus on owning our product adoption metrics. There’s a couple of fundamental questions we have to ask before we sign up for specific OKRs and the KPIs that go along with them.
- Who are we building for?
- What value does the product actually have in the users’ life? And
- How will users discover it?
Insights and definition
Where this starts is usually in a problem space we’re all familiar with, who’s it for, and what is the business impact it could possibly have?
In this insights definition phase, this is where we partner with our PMs, UX research team, and other marketing cross-functional partners, to really dig in and understand what our learning agenda is in actually speaking to customers to understand what they need.
Product requirement document
Once we have our insights defined, and our research initiative out the door, we can come back and actually start a product requirements document. We know who we’re building for, we have some directional insights. Now, let’s put together a document that actually clarifies and crystallises what we’re going to build.
While the PM is writing a PRD, this is often when I kick off my own marketing brief. Oftentimes, some of my marketing brief inputs are inputs to the PRD as well. The two things that come to mind for me when I’m developing a marketing brief are:
- What is the business impact this campaign could have? And
- What is the user need or the market gap this is going to fill?
Go to market strategy
With the PRD and the marketing brief now complete, we’re ready to go to market and of course, that marketing brief will enable creative kickoffs, engaging with our cross-functional partners across the company, it will have outlines for positioning, messaging pillars, all the things you need to develop a strategic campaign.
We are far from done though, once we launch. At Uber, this is especially true. Remember, 700 cities, 65 countries, how are we going to build a campaign that scales outside of London or New York, or Mumbai?
This is where we lean on our regional teams, our operations teams on the ground to equip us with insights that actually help define or refine the positioning and messaging within a campaign on the ground.
Before we are ready to officially grow a product in a market, we often partner with those teams on the ground to playbook our campaign and ensure everyone is speaking the same language about the feature when it lands in the market.
What is an insight at Uber?
Getting back to customer obsession and why it’s so important at Uber, I want to define how we think of insights at the company. There are really three categories in which we are distilling insights from a very wide funnel.
Number one is behavioral. This describes the motivations or behaviors behind something we observed through research.
Example: Instant Pay
An example I can give you is we had a payments product in the United States called Instant Pay, where our drivers can cash out their earnings up to five times per day on a debit card linked to their account.
When I saw the data it reflected drivers were actually hitting this maximum several times a week, I was curious why that could be the case. At Uber, we are famous for championing weekly payments, drivers get their payment on a weekly cadence instead of monthly, like much of Europe or bi-weekly, like North America.
We dug into the data and talked to customers, we understood there are on-platform expenses that need to be taken care of during the course of a day driving on the platform. Petrol, lunch, other miscellaneous car maintenance you have to take care of.
These are things that are really important and not understood through data but understood through conversation. Those are behavioral insights.
Market insights – competitive trends, what’s happening in investment in the space?
Example: Uber Money
An example we looked at recently with the launch of Uber Money last month was the amount of FinTech investing happening across the world, specifically in LATAM. That was a key insight for us to understand there’s actually a market opportunity for us to start offering financial services to a lot of our customers in that market.
Finally, and probably most importantly, is usability. When we put a surface or a feature or an app in front of a customer, how are they engaging with it? Is comprehension high or low? What is the feedback cycle?
These are insights that really drive a lot of our product marketing and product development at Uber.
Ways to build with the customer in mind
I want to share a few tactical recommendations on things I’ve used in the last three and a half years working at Uber that I’ve found really useful as we’ve built out our research functions and marketing functions, that were pretty nascent at the time I joined in 2016.
Develop and socialize a study guide
This one might seem obvious, but anytime you’re entering a new research phase, I find it imperative to develop and socialize a study guide with all the stakeholders who are going to be impacted by the product you’re about to set off and build or the insights you’re about to discover.
This is a great way to unify everyone toward a common objective before you even get to a PRD, before you even get to a marketing brief, or a go to market possibility.
Themes & methodologies
Some of the things that went into one of the study guides I built on the lead up to the Uber Money launch last month and the year preceding it were mapping key objectives around what were the opportunities we were trying to solve for, number one, our customers and then number two, the market.
- How can we get some business growth out of investing in this space?
- What are the themes we already know related to payments experiences on the Uber platform?
- How can we validate those?
- Or how can we pressure test those?
- What methodologies are we going to use in our research; qualitative, quantitative, ethnography?
How are we looking at our data to ensure we’re selecting the right users to participate in a study? Which is arguably the most important thing you can do.
You want to identify a broad set of users to participate, but also make sure you’re selecting the right group that is going to give you pushback on your preconceived notions.
Logistically speaking, who is your translator on the ground? Who is sending out the emails to recruit drivers to join the facility and actually participate at a roundtable with you? Who’s rewarding them?
There are all of these details to think through before you even kick off your first question during a round table.
Finally, the team which I’ll get into in just a moment.
Who’s actually attending on the ground research?
We’re starting to champion more of a cross-functional group attending research around the world, not just product marketers, not UXR, not designers, not PMs, but engineers and data scientists, how do we get everyone involved to get closer to the customer?
Make insight gathering interactive
One thing we’ve done at Uber to make insight gathering and subsequent share outs a little bit more interesting and interactive is leverage tools like Google Plus, RIP, to share out real time insights during research as it’s happening.
This is a great way to engage cross functional teams who aren’t on the ground but still have an interest in what you guys are doing. In 2017, we were rewriting the entire code base for the Uber driver app and we dispatched a team of 11 researchers to 10 different countries over three weeks to get qualitative feedback on the new app rewrite we were putting in front of customers as a beta for the first time.
This is when research at Uber went viral because everyone was participating and so interested to see what was happening in these sessions, that we even had C-level executives posting on Google Plus asking questions.
It was a great engagement tool to get people involved and interested in what we were actually trying to accomplish.
Get in the shoes of your customer
Another thing to do is get in the shoes of your customer. About 150 of us in Amsterdam last summer were signed up to be UberEATS careers. Of the 20 or so trips that I completed, I received zero tips so I need better training before I go out on my next delivery.
But this is a great way to get into the customers shoes and it’s very unique at Uber, you think about the digital world meeting the physical world and the challenges that provides to someone who’s in the moment, who pulls in down a one way street accidentally, who goes to the wrong door to deliver a meal.
There are logistics on the ground you just can’t conquer digitally that need human interaction. These are great ways to to build empathy with the experience you’ve built digitally and then also file a massive number of bug reports that we did that day for the engineering team, because we found a lot of gaps.
Measure twice, cut once
What I mean by this is, we often rest our caps on a range of qualitative insight with a relatively small group of people. I think there’s a lot of value that can come from that, I just think it should expand.
At a company like Uber where scale means everything, I have to validate across a much wider group of customers. When I did qualitative research for the Uber money brand prior to launching in 2019, I found a lot of general support for offering drivers debit cards, finding ways to give our earners on our platform access to their earnings faster.
But I knew that was going to be very different by market. So while I had a nice, diverse qualitative set of findings, I still needed a way to reach more customers, understand nuances by market, which would inform our go to market strategy.
I issued a survey to understand not so much about UX or usability, but more about perception and trust. Uber’s reputation is vastly different market by market. In the US, a driver would say “Absolutely not, I would never use one of those products”. But a driver in India would say “100%, I will sign up the day you launch it”.
There are very, very important cultural nuances to get out of your research and that’s definitely something we do through all the quantitative efforts that we do post-qual.
Build a knowledge base
What do you do once you have a set of insights? How do you make sure all of this customer obsession work is actually being captured somewhere where you can share it, store it, and it doesn’t lose its value over time?
This is where building a knowledge base comes in handy. At Uber, we have teams all around the world conducting their own research and at times, it was very confusing to understand:
- Who had done what,
- Were the insights actually valuable?
- Were they usable?
- Did they talk to the right users?
In a scrappy startup environment, that’ll happen and that’s okay. But find a champion who will build a knowledge base for you, it doesn’t have to be something where you dedicate engineering hours or time to, it can be a Google Drive folder, well organised, of course.
Building a knowledge base is really important to communicate to your internal stakeholders you have a pre existing set of insights to tap into when you’re in the PRD stage, the marketing brief stage, something you can leverage that already exists.
At Uber our fantastic UXR team built a tool called Kaleidoscope – this is a searchable database where you can go in and select topics, customer types, regions, and functions and sort by digestible insights.
This is a hugely valuable tool for me as a product marketer. Again, there can be versions of this that are much smaller, much easier to build, you have to be at a fortune five to get this you can be at a startup and build it in Google Drive as well. But having a single source of truth for where you are actually delivering these insights is really important.
Lead a design sprint
This is typically something reserved for designers or PMs but you’ll see a lot more PMMs at Uber lead design sprints. The great thing is there’s a lot of open source help about how to do this out there in the wild already.
Example #1: Driver Advisory Forum
I had the chance to lead a design sprint with drivers in San Francisco last year at what we call the Driver Advisory Forum. We bring in 40 to 50 drivers from across the United States with very unique needs or time of service on the platform, ratings, earnings, etc. to get a better understanding of how they operate and what’s important to them.
Example #2: Uber Money
Again in the lead up to the Uber Money launch, I sat down and had a financial planning session with a group of drivers and we did this in the design sprint format. One of my favourite exercises was simply having our drivers draw a pie chart: think about your weekly earnings or daily earnings on the Uber platform, and tell me how you spend them.
I think the most jarring one for me working in payments was none of it was being saved, it was all being spent on life’s expenses or operating expenses. That’s where these insights started to hit home and we knew we needed to do more and do it better.
Test, learn, refine, measure, repeat
Oftentimes what customers say isn’t actually what they do in practice. I think a lot of you reading can probably relate to that. When you have a great batch of insights and you’ve built a product and you’ve built a campaign then you go to market and it flops.
Sometimes you wonder why. But how can you rebound? How can you put new messages into market? How can you use different channels to understand what isn’t resonating and what is not?
A lot of times what our UXR counterparts will do, when they see a limited rollout be unsuccessful is actually cold-call customers and say, “You’re interacting with this experience a really funky way what’s going on?”
Surprisingly, they’re very open to giving us feedback and that’s a fantastic resource for us to have. Sometimes, I will go down to what is called the Greenlight hub, which is underneath some of the Uber offices where drivers go to sign up for the platform, get their vehicle inspection checks to make sure they’re good to go.
I’ll just sit in the waiting room and have conversations around certain features that I’m working on or that we’re piloting, to give them a real-time test while they’re waiting to update their registration.
Contextual in-app awareness surveys & sentiment reading
There are all these fantastic avenues to continue testing, learning, refining, measuring the impact, and continuing. Another resource I’ll lean heavily on once I’ve launched a feature is contextual in-app awareness surveys and sentiment reading.
Recently we launched the Uber debit card, we wanted to understand how are you engaging with it? What do you think? Is there anything else we could be offering?
The response rate to these is tremendous and it’s a great service for us to utilize to get learnings as PMMs to take back to our product teams and make improvements.
Develop and socialize a study guide – having a learning agenda and making sure that’s socialized is hugely important. I found a lot of success that way. It makes the research process go a lot smoother, identifying the stakeholders you want involved, who your subjects are, all of those really important details.
Make insight gathering interactive – find channels in which you can distribute insights in real-time, instead of waiting for all of that hard synthesis work and doing one share out around a table.
Measure twice, cut once – validate your qualitative findings in quantitative research.
Build a knowledge base – find a repository in which you can store and reuse insights over time.
Lead a design sprint – find ways to get your cross-functional partners involved early to rally around a common objective before you start building a product.
Finally test, learn, refine, measure – which is a constant loop we’re all intimately familiar with.