Choosing the Right CBSE School A Comprehensive Guide for Parents
Service x berkeley 2015
1. Delivering Real-Time Product Insight
A platform for Product Managers to learn what customers are
saying about their products and competition
Total # of interviews: 139
2. Rob
Rob did marketing
for one of the
most
sophisticated
online marketing
practitioners in
Europe.
He also worked
for Google in
Sales.
He was an NCAA
champion for the
Cal Rugby team
too.
Peter
Peter works with
computational
neural networks.
He also taught
himself Calculus
at age 13.
He enjoys juggling
too.
Anting
Anting built a
language
detection system
using NLP for
Yelp.
He also built
electricals for an
autonomous Mars
rover.
He likes to learn
something of
everything.
Babar
Babar worked
with designers
and developers to
build a platform to
organize group
travel.
He also built a
location-based
Twitter feed for
himself.
He argues about
Soccer Analytics.
3. 3
1. The “Hierarchy of Pain”
2. When to Pivot?
3. Value of Getting Specific
a. How to test a hypothesis
b. Power of customer archetypes
Key Lessons Learned in 10 Weeks
4. Lesson 1: The Hierarchy of Pain
4
Ready to throw money
Attempted a solution
Looking for a solution
Is aware of problem
Has a problem
Where money is ->
(pupils dilating)
Where many
seemingly excited
customers actually are
5. What we used to think...
Customer Support
It’s a massive sector and currently sucks
Team
We’re pretty smart, have strong skills and interesting
technology
Hypothesis
We can automate repeatable queries. Of course customers
will love it...
ServiceX - Week 0
6. 5
1. Large companies
with existing
customer support
teams who want to
increase efficiency.
2. SMBs with little to
no existing customer
service presence
who would like to
establish one at low
cost.
Use NLP and machine
learning to automate
the customer
experience workflow,
to increase the
efficiency of customer
experience consultants
by organizing contact
sources, and to
optimize customer
satisfaction by
providing additional
avenues to reach
support.
1. Conduct customer
interviews with
companies.
2. Train a model on
customer support
conversation data.
3. Build prototype.
1. Companies with
existing customer
support teams who
want to increase
efficiency.
2. Potentially other
providers in the
space with
infrastructure we
can leverage/act as
a complement to
1. Customer acquisition.
2. Computing.
SaaS subscriptions: charge companies based on
their size.
Initially, social media,
specifically Twitter.
Other platforms and
other channels can
follow later.
Dataset of customer
service interactions
(need quality and
quantity).
1
3
4 2
We provide a SaaS
product that is
standalone or that
interfaces with an
existing CRM.
6
7
89
Business Model Canvas in Week 1
7. Week One MVP: Automating Hootsuite
7
MVP Week 1: Supercharging Social
8. “Time spent answering repeated queries is painful
but Zendesk has some pretty neat DIY functionality
that we have been using as a ‘Hack’ to mitigate this”
- Stacey, Customer Support Manager at Consumer
Application with > 50MAU
8
MVP Week 1: Supercharging Social
9. Lesson 1: The Hierarchy of Pain
9
Ready to throw money
Attempted a solution
Looking for a solution
Is aware of problem
Has a problem
Where money is ->
(pupils dilating)
Where many
seemingly excited
customers actually are
18. 18
UX researcher within the product group at a
leading consumer drone manufacturer
To learn about her customers, Jenny:
- Reads drone hobbyist forums online
- Prepares customer surveys and collate results
- Prepares weekly reports for the company, with
key customer metrics and customer quotes
- Reports to her boss “Jeremy” with actionable summaries of customer
feedback and metrics
Customer Archetype - “Jenny”
19. The Net Promoter Score (NPS)
NPS is the gold-standard of pre-social media times
But NPS is still the metric that Jenny cares most about.
Companies cannot calculate real-time NPS, and struggle to gather statistically
significant results.