Professional Documents
Culture Documents
KNOWLEDGE
IS PRODUCT
CARLOS GONZALEZ DE
VILLAUMBROSIA
Founder & CEO
The world just doesn’t work like that. Data alone won’t get you far. What you
need is product analytics. This is the structured process through which you
not only acquire data, but also use it to extract meaningful insights to guide
your decision making.
The Product Analytics Economy looks at how Product Managers today are
using product data for innovation, iteration and inspiration. With ample in-
sights provided by Heap, a leading digital experience insights provider, you’ll
learn how companies on the cutting edge make data work for them.
Inside, you’ll find insight from Nike, YouTube, Heap, Disney, Waze, Pipe-
drive, Bumble & Badoo and more. You’ll explore real life examples of how
leading firms are harnessing the potential of data to facilitate the transition
to Product-Led Growth, and find yourself liberated from some of the many
misconceptions that surround the use and application of data today.
Data is at the center of everything we Product Managers do, so let’s learn how
to make the most of it! Knowledge is power, but only when it is applied.
5
CHAPTER 01
HOW TO
MAKE
PRODUCT
ANALYTICS
USELESS
6
MATIN MOVASSATE
Founder & Chairman
To illustrate this point, imagine that January 1 rolls around, and you decide,
like many others, to start exercising. But first, you need to research and Matin
Movassate is the co-founder of Heap, a well known data analytics platform.
So why is he interested in telling us how to make Product Analytics useless?
Matin’s thesis can be summarized in one phrase:
To illustrate this point, imagine that January 1 rolls around, and you decide,
like many others, to start exercising. But first, you need to research and join
an appropriate gym. This takes some time and research. Then you need the
right outfit to work out in., Ok, after a couple of weeks you’ve done that. Now,
you need to determine what kind of exercise will best suit your goals.… and by
now weeks have passed, it’s already mid-February, and your motivation is
not what it was! Best just wait until next year, right?
This metaphor demonstrates the common fallacy that having the thing is the
same as doing the thing. This is not true. And while we all know that this is
not true, we often don’t act accordingly. In the Product world, argues Matin
Movassate, data companies pitch that merely having data is the key to ma-
8
CHAPTER 01
king a better product. But, just having the data is not the same as using it to
make better decisions.
To not be useless, Product Analytics should connect the dots between the
data you have and the actual decisions you need to make to deliver something
amazing for your users.
So, how does he recommend doing this? Three key pieces of advice
"There is a disproportionate payoff from the unknown, since you typically have
little to lose and plenty to gain." Nicholas Nasim Taleb, The Black Swan.
Search for impact. Ask a hundred product leaders what keeps them up
at night, and 99 will say “prioritization.” (The remaining leader is too
oblivious to know.) How to avoid prioritizing the wrong things? The main
thing is to set baseline KPIs for all parts of the product, both individual
user flows and larger adoption and engagement metrics. The best way
to set these baseline KPIs is to gather reliable usage data (not unreliable
data - see point #1!), then ruthlessly track engagement numbers against
business goals. Be explicit about what you’re not prioritizing and why.
Then submit the whole prioritization structure to the next point...
Hypothesize and iterate. The best way to figure out what to do is to learn.
The best way to learn is to make hypotheses and roll out an experiment.
9
This process forces you to measure effectiveness against success criteria,
and adopt an experimental mindset. When every new initiative is unders-
tood as an opportunity to learn, then it becomes easier to make incremen-
tal changes because you understand that not all changes will be “success-
ful” - but they will all lead to important lessons that, over time, will allow
you to move your product in the right direction. This attitude is just as
important for fixing the product directly as it is for deciding what areas in
the product to focus on.
“The value of a hypothesis is how much signal it gives you, not necessarily
whether it’s right or wrong.” - Matin Movassate
No product is ever perfect. All Product Managers strive to improve their pro-
ducts over time. Data is essential for this, but it won’t get you there alone. To
improve, use data combined with a smart, tested strategy that will, over time,
get you, your team and your product moving in the right direction. Tools ma-
tter, but only if you use them!
10
11
CHAPTER 02
PRODUCT
ANALYTICS
MADE SIM-
PLE
12
ENZO AVIGO
Co-founder and CEO
Product Analytics may seem difficult or inaccessible to some with the satura-
tion of available data and noise in the industry. However, with a clear goal set
in place, achieving the right analytics stack and ramping up your skills can be
quickly achieved.
In this chapter Enzo Avigo, former PM at Intercom and Co-founder and CEO
at June.so, proposes a framework of simple steps to help Product Managers
finally cut the noise and uncover invaluable insights about their features or
products.
We’ll look at a four step process for getting started with Product Analytics
without getting overwhelmed, and then drill down into some of the most
useful metrics you should be keeping an eye out for right from the start.
Start lean. You need to avoid heavy lifting with data projects. A typical
journey with analytics begins by setting up some success metrics, then
tracking user behaviors, which is called instrumentation.
Usually people tend to go overboard and define too many success metrics,
and then instrument a lot of things in their product, and analyze them.
This method ends up taking too long for any progress to happen. Enzo
recommends you avoid that approach and instead take on only one project
at a time and you take it along that entire journey.
14
CHAPTER 02
Say you have a new feature you want to release. Start with defining a suc-
cess metric, then instrument two or three behaviors related to that feature
followed by tweaking around your favorite analytics visualization tool to
lend the analysis result that you're seeking.
Set your goals. When defining your goals you need to be very pragmatic
and break things down. So a goal is typically a high level achievement that
you want to drive. It will usually look something like: I want more cus-
tomers, or I want customers to spend more. You can further break these
goals into subgoals which tend to be the starting point of the project you
select to start your analysis journey mentioned above.
Seek best practices. Make sure to utilize online resources to your benefit.
You can start by talking to people around you, to heads of product, CEOs
of companies you benchmark and get the conversation flowing.
Align on mindset. Finally, you need to make sure your team is also alig-
ned with your data analytics mindset. Enzo refers to this as the integra-
tion process and ensuring it matches your company culture. This can be
done by trying to fit this data mindset into your day to day - and make it
fun!
Now that we've seen the processes to get started, let's look into the metrics
relevant for early stage companies or for your product team:
Active users. These are key to helping you reach your product market fit.
A good way to measure that is by tracking when users are active, meaning
when they're getting the unique value from your product.
15
Feature activation. The most granular way you can get started is to un-
derstand whether or not people engage with specific features. Track the
first steps of a feature being used and use a basic funnel to understand,
from the start of your experience, how many of the people are actually
engaging at least once with one of your specific events.
Feature retention. This is basically how long people stay around on your
product. You just need that through that in a, you know, basic line curve
analysis and understand how long, how often people like are sticking to
these features or products over time.
Now you know how to get started with Product Analytics without getting
overwhelmed!
16
17
CHAPTER 03
DISCOVER
A PRO-
DUCT FO-
CUSED FRA-
MEWORK
FOR MEASU-
RING SUC-
CESS
While OKRs Are Popular, They Have Their
Limitations. Here’s an Alternative.
18
RAPHA COHEN
CPO at Waze
No evaluation of whether or not you are capable of hitting your Key Re-
sults
Putting context at the center of your Product Management approach can help
you better understand your users in a more deep, practical and profound
manner. Questions to help with this are:
20
CHAPTER 03
Understanding the worldview of users of Waze allowed Rapha and his team
to better reach their KPIs of acquiring more miles per session. They did this
not by focusing only on the data, but by placing this data in the context of the
user's life, and understanding other things that may have motivated them,
such as altruism or fear of rejection.
A useful system used by Google and others when building products or fea-
tures is: Goals, Signals, Metrics, this can help you understand the Objective,
Signals, Key Results aspects of HOSKR
Goals: How do you want your users to engage with your product?
Signals: What user behavior would indicate your goal has been met?
21
Imagine HOSKR as a Loop
Your Hypothesis allows you to determine the Signals you will observe that
inform you as to whether or not you’re achieving your Key Results on the way
to your Objective. As you observe more Signals and gather more data, you
will be able to revisit your initial hypothesis, updating it accordingly as your
company grows, your user base evolves, and you gain more insight.
Here’s how a company like Google Waze could look at carpooling if they were
using OKRs:
Observation: 65% of carpool rides are with “exes” (former user pairs)
Data: Most users can’t easily find their “exes” in the app
Key Result: Increase share of rides with exes from 65% to 85%
22
CHAPTER 03
Objective:
Driver centric: Fully utilize my car and save more money
Rider centric: Relax and book my commute
Business centric: We can monetize this feature
Signal: Only a fraction of engaged riders are interested. Those set rides
before getting incoming requests. When such a ride is requested, riders
flow in masses
The OKR objective to maximize riders with “exes” may, as shown by the
HOSKR framework, actually resulted in fewer users of the app as it failed
to take into account the users worldview. Instead, the pre-approval option
achieved more benefit with less downside.
Conclusion
“Form a cohesive view of the interactions between users actions and the universe”
23
CHAPTER 04
HOW
YOUTUBE IS
PREPA-
RING TO
ONBOARD
ONE BI-
LLION MORE
USERS
Learn the Prioritization Secrets of
YouTube’s Head of Product.
24
MINAL MEHTA
Head of Product at Youtube
Despite YouTube’s popularity, Minal works with a very lean team. This means
prioritization is critical. But this isn’t only true for YouTube: Many of us
now work with distributed teams, perhaps for the first time, and always are
juggling various goals and commitments. Prioritization matters now, perhaps
more than ever.
Understanding the actual implications of what you are really saying yes to is
the key to prioritizing effectively. Something may sound simple on the surfa-
ce, but could it lead you down a rabbit-hole that will deprive you of your time,
energy, and resources?
26
CHAPTER 04
When you have a strong understanding of both your current users and the
new users who you anticipate will be joining your service, then intelligent
prioritization can follow naturally from knowledge of your users. If you know
who your expected users are and what matters to them, then you can make
resource allocation decisions accordingly.
All Product Managers deal with complexity all the time. Minal Mehta is an
advocate of the Cynefin system which can help you find clarity in both simple
and complex contexts.
27
Minal’s Simple Rules for Work and Life
Detach. Minal works from 9 to 6, with a quick email download at 9pm. She
ensures work and personal phones are not on the nightstand!
Make Time for the Important: Block out 2 hours a day for non-reactive
work
Manage teams: Communicate early even when you don’t have all the
answers. Take simple one on one standing meetings. Don’t multitask! Let
people resolve issues in their own time i.e asynchronously
Everyone is different, and finding your own rhythm for prioritization and
work life balance will depend a lot on who you are, what matters to you, and
what works for you. You must know your users. You must also know yourself!
28
CHAPTER 04
29
CHAPTER 05
HOW TO
MAKE THE
TRANSITION
TO
PRODUCT-LED
GROWTH
30
RACHEL OBSTLER
VP of Product
Your product is your primary support tool and primary marketing tool
Your model is scalable - your ability to acquire new customers is not limi-
ted by the number of calls your sales team can make.
If you’re thinking about making the transition to Product-Led, then Rachel
has some advice to help you navigate the challenges you will encounter along
the way:
32
CHAPTER 05
Example.
33
Solution. Understand the difference between growth and sales metrics.
Determine the metrics that are important for growth, and optimize for
them. Adopt a bottom-up strategy. If your bottom-up strategy is success-
ful, sales can add value by taking the momentum higher up in the organi-
zation.
Example. Heap helped its client, Amway, use autocapture to make their
data analysis flexible so they could iterate quickly instead of having to
wait months for tagging. By removing the friction and making the value of
analytics easily accessible, Always was able to roll out Heap’s tool to over
200 of their employees.
34
CHAPTER 05
PLG presents huge opportunities for growth. Adopt it before your competitors do!
35
CHAPTER 06
PRODUCT
STRATEGY:
YOUR COM-
PETITION IS
NOT WHO
YOU THINK
IT IS
Krishna illustrates a key “red pill moment” that changed his perspective on
competition from a Product Management perspective with a story about his
time at the early days of Skype, back in 2009.
Back in 2009, Krishna was an eager PM working in the early days of Skype.
They shipped many features such as screen sharing, Skype Wifie, and video
messaging, and were looking at 50% year on year growth.
The iPhone was out but it wasn’t a big part of the Skype team’s focus. At first,
it was dismissed by many as a “bad phone.” But, of course, it would go on to
transform every aspect of how people communicate with each other.
38
CHAPTER 06
You see, 2009 was also the year that WhatsApp launched. Right now, What-
sApp are almost at 2 billion active users! Skype, on the other hand, went
through something of a rise and fall.
The smartphone era had arrived, and Skype was going to lose badly to an app
- something it had not previously even considered as competition.
Ready-made for the smartphone era, WhatsApp had three big advantages
that Skype did not:
Simple Interface. The instant messaging experience used to be a very di-
fferent and more complex thing! What’s made it as simple as “free SMS.”
The guiding principle that laid this successful path for WhatsApp: “How do
we reduce time for first conversation to (almost) nothing”
Fun fact: In 2009, Skype had 1000 people to WhatsApps 55!
The Paradigm Shift: Your competition is not always who (or what) you think
it is.
“Startups mostly don’t compete against each other, they compete against no one
giving a shit” - Justin Kan, founder of Twitch.
39
Skype didn’t realize they were competing with the frictionless experience of
an app built not for computers, but for smartphones. Perhaps your real com-
petition is also different to what you think it is.
It may not be a product. You may be competing with apathy. With the status
quo. As Krisha explains, it is important to start thinking about competition in
a much broader sense.
As the CEO of Netflix, Reed Hastings, says, they are competing against
“Drinking a bottle of wine” - Netflix views their competition as any other
way of spending your evening not on Netflix. They weren’t focused on
Blockbuster. They were focused on you.
“What business are we in.” (The answer won’t always be as simple as you
expect).
“Who or what are we actually competing with.”
40
THINK OF ANY
NIGHT YOU DID NOT
WATCH NETFLIX... WE
COMPETE WITH DRIN-
KING A BOTTLE OF
WINE. THAT’S A PARTI-
CULARLY TOUGH ONE.
CHAPTER 07
WHEN
LOSING A
USER GETS
YOU CLO-
SER TO
YOUR COM-
PANY MIS-
SION
Customer Acquisition Matters, But Not at
All Costs.
42
MILES NORRIS
CPO at Bumble & Badoo
As VP of Product, Miles and his team went on a two year journey to unders-
tand what happens when the KPIs you are given from the business team, and
the core mission of your business, are in conflict.
44
CHAPTER 07
Early “Aha!” moments. So, they worked to get every user a match within
24 hours.
Customized experience - the product team added more tools for the user
to make their experience unique
Results. Activity distribution. Some users had a lot of activity, others almost
none. More MAU growth. Still no meaningful impact on M1 retention.
Retention remains a key KPI
45
relationships. Since the actual dates take place off the app, they had to find a
proxy metric for this. They settled on:
Good chats. The product team improved the in-app chat experience with
stickers, GIFs, etc. This improved the frequency of use of the chat function.
If users leave, then why? They asked users for deletion reasons, and that’s
when the Aha! Moment finally came: OH - it’s because they had succee-
ded and found a relationship!
So, what are the “bad churn” reasons someone would leave their app?
They broke this into two categories:
46
CHAPTER 07
Core Learnings
After a two year journey with retention as a goal, the MagicLab product team
realized the goal was not that simple. The core retention metrics not working
and distracting the team from the leavers they pulled. It’s important to
always ask these questions:
Questions
Are your metrics correct?
Sometimes, retention really may not be the goal you’re looking for!
47
CHAPTER 08
HOW TO
BUILD DATA
DRIVEN
ROADMAPS
48
SUSHMA RAO
Lead Product Manager
Sushma Rao, Lead Product Manager at Nike, has worked in everything from
architecture to aerospace engineering, a blend of experiences that has led her
to the Product Roadmap ideology she has today.
Sushma makes her roadmaps with a combination of data, instinct, and com-
munication. She has a five-step process to creating products, couched within
an agile mindset; as she moves from one step to the next she and her team
are constantly iterating and questioning their product rationale.
50
CHAPTER 08
This realignment from the bottom up is possible, but needs data to back it up,
which leads us to the next step in the product creation process.
Data.gov
If you haven’t established metrics at this point in the process, this is the time
to do it. The sooner you can define your product metrics, the better, because
they’ll inform all of your decisions moving forward. Tie metrics to strategic
goals, and if the data shows you need to adjust your strategy, do it. And if
you are losing your way in big data, narrow down the data you collect only to
those which are relevant to your industry or roadmap.
51
If you’re creating a UI-based product, for example, you’d focus on engage-
ment and usage metrics like product adoption or actions taken. But if you’re
working on a SaaS-based product, you might focus on customer acquisition
cost or lifetime value.
This is where you confirm that it’s viable for your company and your team
to build this product. Does it make sense for the business? Is it doable within
your timeline and with your team? Estimate the economic value and custo-
mer lifetime value of the product or feature if you were to fully build it. Your
end-goal shouldn’t just be revenue, but it’s helpful to look at revenue as a
metric to make sure your product aligns with overall business goals.
But now we’re in the launch phase, where more complicating factors enter.
Perhaps counterintuitively, there are a lot of roadblocks during this time
because you’re finally taking action on all of your thinking and building. And
taking action means there are more external factors to take into considera-
tion, and new things to measure.
It’s no longer just the solutions team working on the product; now you have
to align with other teams like tech support or operations. This creates contin-
gencies and can muddle up your product direction.
52
At this stage, use data as a tool to help you manage priorities. Find the fra-
mework that works for you, like WSJF (weighted shortest job first) or RICE
(reach, impact, effort, confidence). Use a combination of data analysis and
gut instinct to organize your priorities into a prioritization framework, and
move forward keeping data metrics and other teams' roadmaps in mind.
make space for different individual values within the overarching values of
the company. This is why giving people pre-make solutions for big problems
seldom works. You have to let them honor their own personal values at work,
and thus allow them to grow into themselves.
Nike implemented these feedback loops in a few instances, which led to pro-
ducts like Nike Live, Buy-Online Pickup In-Store, and Reserve Lockers. The
common-sense wisdom of that time was that brick-and-mortar stores were
losing popularity, so it would be a mistake to invest in in-store features. But
Nike’s feedback loop data showed them that there was in fact a high demand
for in-person experiences, but for non-traditional and lower-friction ones.
Conclusion
Metrics are essential, but be careful of blind faith in data. Think like a statis-
tician; if your results don’t make sense, remember that correlation does not
equal causality and try approaching it from a different angle. And also, me-
trics are for sharing! Involve your team for better alignment and understan-
ding of the roadmap. Create your metrics with your team, and then measure
and learn from them together to keep building better and better products.
CHAPTER 09
HOW ME-
DIA COM-
PANIES ARE
MANAGING
THE DIGITAL
TRANSFOR-
MATION
56
CHAPTER 09
57
Old Media Opportunities
Within the Old Media world, there are now several opportunities presented by
the overlapping fields of Product Management and tech:
Sending their existing high value content to Streaming Video on Demand
(SVoD).
“If you ask five different executives at media companies what their digital strategy
is, you’ll get 7 different answers.” - Rachel Bailin
58
CHAPTER 09
New Tech
New tech companies are firms like Netflix and HBO who are at home in the
world of digital streaming and tend to put technology first, focusing on op-
timizing individual user experience. Like Old Media, new tech is currently at
an inflection point and is facing various challenges and opportunities which
could engage a Product Manager.
Cultivating Brand. Does the fact that a show is created by Netflix imme-
diately make you want to watch it?
Risk Taking. A focus on optimization can lead to less creativity and more
bland, cookie-cutter content. Viewers need to be surprised!
59
CHAPTER 10
HOW TO
GENERA-
TE INSI-
GHTS AND
INFLUENCE
PRODUCTS
60
MATIN MOVASSATE
Founder & Chairman
We began this book with the provocative claim about how to make Product
Analytics useless. A bold claim indeed, especially coming from an analytics
company like Heap! But the point is this: Product data is in fact enormously
valuable, it just takes a little bit of work to make your analytics tools yield
results.
Many tool companies out there will try to sell you on the lie that simply
having this data is enough to build great products. All you have to do is hand
your teams a string of numbers and POOF out pops a world-changing
product.
So if having a great dashboard is not the goal of your analytic setup, what is?
The primary way that data produces amazing products is through the insi-
ghts it provides. By understanding what an insight is, and thinking carefully
about how you acquire them, you can work to structure your product develo-
pment processes around getting more of them.
What Is An Insight?
Everyone Talks About Insights! But What Are They… And How Can You Get More
From Your Data?
Going back to basics, Heap CEO Matin Movassate shares the definition of
‘insight’ that he’s been working on, as it applies to Product Management:
62
CHAPTER 10
WHAT IS AN INSIGHT
Everyone has a mental model, no matter how vague, about what their product
is, what it does, how and who uses it, etc. Over time your story should beco-
me more detailed, as you learn more about your users. It should also become
more sophisticated. For example, understanding how different segments
of your users behave differently. An insight is a piece of information that
changes your presumptions or suppositions about your product. It changes
your mental model and your story. Perhaps you’ll learn that something you
thought was easy is actually pretty hard, or that people use your product in a
slightly different way.
Real world example: Sur La Table found that product page views correlated
highly with purchases. So they drove usage to product pages, which increased
their overall purchase rate by 6%.
63
1. Use tools that surface the unexpected
The ideal tool should be one that shows these to you automatically, so you
don’t have to find them yourself. We call them “proactive insights,” but
you can call them “what users really do.”
If you’re looking to change your story, you need to run experiments. You
can imagine the story you tell yourself about your product as the status
quo. As a scientist, you need to challenge the status quo with a hypothesis.
Let’s say you run a SaaS site, and you write a blog series about how to use
your new feature. Your hypothesis is that releasing this blog series will
increase your adoption rate. You can test this by seeing how many people
who read the blog went on to adopt the new feature, compared to those
who did not.
Challenge Measure
assumptions and test
64
CHAPTER 10
The story of the genius who wakes up one morning with their whole
product planned to perfection is nice, but we’ve never seen it in real life.
Great products are built by a series of small improvements. You chip away
at your product, making adjustments and changes. Maybe you chip away
and reveal something huge that you never anticipated.
These small changes are what allow you to gather insights in the first
place. You make hypotheses, you roll out small changes, and you see how
those changes are adopted by the market. Small changes open the door to
bigger advancements.
As you continue to build and iterate on your products, the future of digital
product insights is looking brighter. With new product capabilities, products
like Heap are powering automated insights and removing barriers to success.
Reducing hours spent parsing through data for product and analytics teams, a
real digital experience insights platform will empower you to spend more time
building the right product.
LEARN MORE
65
THE
PRODUCT ANALYTICS
ECONOMY
There has never been a better time in history to build digital
products. At the same time, customers' expectations are
higher than ever before. This means that delivering an
exceptional digital experience is the only way to earn trust
and drive revenue growth.
Matin Movassate, Founder & Chairman at Heap Krishna Panicker, VP of Product at Pipedrive
Minal Mehta, Head of Product at Youtube Rachel Bailin, former Director of Product at Disney