1
00:00:06,600 --> 00:00:07,950
- Now let's talk about

2
00:00:07,950 --> 00:00:10,470
defining customer research techniques.

3
00:00:10,470 --> 00:00:12,360
If the key to building the right product

4
00:00:12,360 --> 00:00:14,310
is to understand customer needs,

5
00:00:14,310 --> 00:00:17,370
we need to rely on customer research.

6
00:00:17,370 --> 00:00:20,160
However, in many instances,
our understanding

7
00:00:20,160 --> 00:00:23,520
of what the customers need,
or even who the customers are

8
00:00:23,520 --> 00:00:25,200
is not accurate.

9
00:00:25,200 --> 00:00:27,750
No matter how well we
know those customers,

10
00:00:27,750 --> 00:00:31,680
our understanding of what they
need is just an assumption

11
00:00:31,680 --> 00:00:34,530
unless we create the right data.

12
00:00:34,530 --> 00:00:38,670
This is a hypothesis and we
need data to have it validated.

13
00:00:38,670 --> 00:00:41,587
As Henry Ford is said to have stated once,

14
00:00:41,587 --> 00:00:44,790
"If I had asked people what they wanted,

15
00:00:44,790 --> 00:00:47,610
they'd have said a faster horse."

16
00:00:47,610 --> 00:00:50,160
This way a car would never be invented.

17
00:00:50,160 --> 00:00:52,650
A true product visionary frequently comes

18
00:00:52,650 --> 00:00:56,610
with the ideas that others have
not yet suggested or tested.

19
00:00:56,610 --> 00:00:59,820
And yet there is a need
to validate those ideas

20
00:00:59,820 --> 00:01:02,193
before implementing them commercially.

21
00:01:03,180 --> 00:01:06,900
User research focusing on
understanding user behaviors,

22
00:01:06,900 --> 00:01:10,950
needs, and observations
through a number of techniques

23
00:01:10,950 --> 00:01:12,750
is really important.

24
00:01:12,750 --> 00:01:16,350
Those techniques range
from the observation

25
00:01:16,350 --> 00:01:18,630
of users in their natural surroundings

26
00:01:18,630 --> 00:01:21,360
to interviewing them about the experience

27
00:01:21,360 --> 00:01:25,170
with a specific working product
and everything in between.

28
00:01:25,170 --> 00:01:28,590
It's an iterative and cyclical process.

29
00:01:28,590 --> 00:01:31,350
In lesson two, we discussed
the lean startup framework

30
00:01:31,350 --> 00:01:34,200
that provides ways to
validate customer hypothesis.

31
00:01:34,200 --> 00:01:35,033
And we spoke

32
00:01:35,033 --> 00:01:37,890
about design thinking models
that are used to design

33
00:01:37,890 --> 00:01:39,360
and prototype products.

34
00:01:39,360 --> 00:01:44,073
However, this is not sufficient
to define the right product.

35
00:01:45,570 --> 00:01:48,600
In order to achieve validated
learning from customers,

36
00:01:48,600 --> 00:01:53,010
four steps as regards to user
research are most important.

37
00:01:53,010 --> 00:01:56,400
The first one is to
identify the right customers

38
00:01:56,400 --> 00:02:00,540
that represent the target
personas for your product.

39
00:02:00,540 --> 00:02:02,730
The second one is to define the way

40
00:02:02,730 --> 00:02:04,410
of collecting this feedback.

41
00:02:04,410 --> 00:02:05,850
Could be customer interviews,

42
00:02:05,850 --> 00:02:08,910
observation of customer
interaction with prototypes

43
00:02:08,910 --> 00:02:13,260
or existing products, surveys,
web analytics, and so forth.

44
00:02:13,260 --> 00:02:16,440
Third one is to actually
conduct this research

45
00:02:16,440 --> 00:02:20,910
and use measurable results
to validate your hypothesis.

46
00:02:20,910 --> 00:02:23,400
And finally, you have to aggregate

47
00:02:23,400 --> 00:02:27,480
and interpret all the data from
the research and compare it

48
00:02:27,480 --> 00:02:32,430
against the hypothesis in a
non ambiguous, data-driven way

49
00:02:32,430 --> 00:02:36,690
and do your validation
and the final decision.

50
00:02:36,690 --> 00:02:39,240
So once the problem is identified

51
00:02:39,240 --> 00:02:42,840
it is very important to
define how to measure it

52
00:02:42,840 --> 00:02:46,950
and how many respondents are
required for your research.

53
00:02:46,950 --> 00:02:49,200
For example, if the user research method

54
00:02:49,200 --> 00:02:51,630
involves comprehensive
interviews, then the number

55
00:02:51,630 --> 00:02:56,220
of customers who participates
in the research is fairly low.

56
00:02:56,220 --> 00:02:59,100
And if you do web analytics research,

57
00:02:59,100 --> 00:03:01,530
then the number is in hundreds.

58
00:03:01,530 --> 00:03:05,430
Finally, it is helpful to
identify success criteria.

59
00:03:05,430 --> 00:03:07,650
Whether your hypothesis will be validated

60
00:03:07,650 --> 00:03:10,140
by a simple majority of respondents

61
00:03:10,140 --> 00:03:13,050
or you need a specific percentage of users

62
00:03:13,050 --> 00:03:15,270
that will confirm your assumption.

63
00:03:15,270 --> 00:03:18,030
And the most important rule

64
00:03:18,030 --> 00:03:20,520
and the mistake that
frequently is being done

65
00:03:20,520 --> 00:03:24,750
is to never ask a question
that starts with, would you,

66
00:03:24,750 --> 00:03:26,190
would you buy it?

67
00:03:26,190 --> 00:03:27,360
Would you use it?

68
00:03:27,360 --> 00:03:30,960
Those responses do not
validate customer behavior.

69
00:03:30,960 --> 00:03:33,660
And actually most customers would confirm

70
00:03:33,660 --> 00:03:36,000
that they're interested in using

71
00:03:36,000 --> 00:03:38,790
the tool or software or product

72
00:03:38,790 --> 00:03:41,490
because there is no
obligation on their part yet.

73
00:03:41,490 --> 00:03:44,760
In addition, these types of
questions cause respondents

74
00:03:44,760 --> 00:03:46,920
to guess about behavior

75
00:03:46,920 --> 00:03:50,070
rather than share facts
about their actual behavior.

76
00:03:50,070 --> 00:03:52,680
You can show your customers
a prototype of your product

77
00:03:52,680 --> 00:03:54,390
followed by a set of questions,

78
00:03:54,390 --> 00:03:57,960
or simply observe the
interaction with the prototype.

79
00:03:57,960 --> 00:04:00,060
That's much more impactful.

80
00:04:00,060 --> 00:04:02,850
And finally, you can
always conduct interviews.

81
00:04:02,850 --> 00:04:06,090
A simple and yet very
powerful interviewing approach

82
00:04:06,090 --> 00:04:08,610
is called a three point interview.

83
00:04:08,610 --> 00:04:11,310
It consists of three questions.

84
00:04:11,310 --> 00:04:13,440
Question number one, have you ever had

85
00:04:13,440 --> 00:04:15,090
like this specific problem?

86
00:04:15,090 --> 00:04:16,226
If they say no at this point,

87
00:04:16,226 --> 00:04:20,460
assumption is invalidated
and you have to pivot.

88
00:04:20,460 --> 00:04:23,250
Start with something totally different.

89
00:04:23,250 --> 00:04:24,540
Second question.

90
00:04:24,540 --> 00:04:25,373
Tell me more

91
00:04:25,373 --> 00:04:28,200
about the last time you
had this specific problem.

92
00:04:28,200 --> 00:04:29,400
What did you do?

93
00:04:29,400 --> 00:04:33,240
Here you look for them to
actually do something about it.

94
00:04:33,240 --> 00:04:35,190
And third question for you.

95
00:04:35,190 --> 00:04:38,160
What is the ideal solution
in that situation?

96
00:04:38,160 --> 00:04:40,080
So third question is not perfect,

97
00:04:40,080 --> 00:04:42,450
and yet it gives you the direction

98
00:04:42,450 --> 00:04:45,633
for your solution that you
can at least start with.

99
00:04:46,590 --> 00:04:48,270
So now let's talk a little bit

100
00:04:48,270 --> 00:04:50,790
about conducting the research itself.

101
00:04:50,790 --> 00:04:53,070
Customers have to be available first

102
00:04:53,070 --> 00:04:54,720
and in agreement with the research.

103
00:04:54,720 --> 00:04:57,750
Sometimes they have to sign an agreement.

104
00:04:57,750 --> 00:05:00,060
No matter whether they're
internal or external

105
00:05:00,060 --> 00:05:02,250
they need to be recruited.

106
00:05:02,250 --> 00:05:04,830
And there are multiple ways
of reaching out to customers

107
00:05:04,830 --> 00:05:07,860
and giving them different
sorts of incentives

108
00:05:07,860 --> 00:05:10,710
so that they participate in your research.

109
00:05:10,710 --> 00:05:12,720
For example, in case of external customers

110
00:05:12,720 --> 00:05:16,380
you can offer them participation
in a better product

111
00:05:16,380 --> 00:05:20,280
or discount on your product
for a specific period of time.

112
00:05:20,280 --> 00:05:22,230
I give them a gift card.

113
00:05:22,230 --> 00:05:24,570
In case of internal customers,

114
00:05:24,570 --> 00:05:26,940
you can incentivize
them by supporting them

115
00:05:26,940 --> 00:05:29,733
and solving the actual
problem that they're facing.

116
00:05:31,170 --> 00:05:34,950
Also, researchers have to
have an interview script.

117
00:05:34,950 --> 00:05:37,680
They have to allocate
sufficient time for interviews.

118
00:05:37,680 --> 00:05:39,240
Never rush an interview.

119
00:05:39,240 --> 00:05:43,440
If you are absolutely time
boxed, just cut it short,

120
00:05:43,440 --> 00:05:45,900
don't try to rush through answers.

121
00:05:45,900 --> 00:05:48,240
You have to make your
participants comfortable

122
00:05:48,240 --> 00:05:50,640
and very clear about the expectations.

123
00:05:50,640 --> 00:05:54,840
And if there are any legal
aspects and other constraints,

124
00:05:54,840 --> 00:05:56,850
you have to take care of them in advance.

125
00:05:56,850 --> 00:05:59,400
For example, if you may want

126
00:05:59,400 --> 00:06:03,780
to have your external participants

127
00:06:03,780 --> 00:06:06,930
to sign a nondisclosure agreement or NDA

128
00:06:06,930 --> 00:06:10,530
before showing them any
prototypes of your future product.

129
00:06:10,530 --> 00:06:14,310
And all the results are accurately
documented by interviews

130
00:06:14,310 --> 00:06:15,963
in a predefined format.

131
00:06:17,310 --> 00:06:20,550
Tips for conducting user
interviews are as follows.

132
00:06:20,550 --> 00:06:24,060
The most effective interviews
are with one person at a time

133
00:06:24,060 --> 00:06:27,690
because groupthink, when one
answer influences others,

134
00:06:27,690 --> 00:06:30,600
is misleading and happens frequently.

135
00:06:30,600 --> 00:06:33,480
Also when there are multiple researchers

136
00:06:33,480 --> 00:06:35,700
the respondents may feel uncomfortable

137
00:06:35,700 --> 00:06:38,280
so try to do it one on one.

138
00:06:38,280 --> 00:06:40,110
It is important to know the goals

139
00:06:40,110 --> 00:06:41,610
and questions ahead of time.

140
00:06:41,610 --> 00:06:44,460
So give this courtesy to your interviewees

141
00:06:44,460 --> 00:06:46,200
if it is possible.

142
00:06:46,200 --> 00:06:48,360
A prioritized list of assumptions

143
00:06:48,360 --> 00:06:50,490
is used to generate a list of questions.

144
00:06:50,490 --> 00:06:52,200
You don't share this list

145
00:06:52,200 --> 00:06:56,460
but it's helpful to have
those with among interviewers.

146
00:06:56,460 --> 00:06:58,830
It's also advisable to clarify

147
00:06:58,830 --> 00:07:02,430
and adjust the questions
based on the respondents.

148
00:07:02,430 --> 00:07:06,000
But again, there always
has to be a clear script

149
00:07:06,000 --> 00:07:07,053
to start with.

150
00:07:08,550 --> 00:07:12,450
Now you have to define
what interview type it is.

151
00:07:12,450 --> 00:07:14,670
Try to separate behavior interviews

152
00:07:14,670 --> 00:07:16,560
from feedback interviews.

153
00:07:16,560 --> 00:07:17,393
If the goal

154
00:07:17,393 --> 00:07:20,220
for the research is to
understand customer needs,

155
00:07:20,220 --> 00:07:21,510
then ask about the gaps

156
00:07:21,510 --> 00:07:24,810
between their needs and the
available functionality.

157
00:07:24,810 --> 00:07:27,510
If the goal is to get
the feedback on features

158
00:07:27,510 --> 00:07:30,840
of a specific product, use your prototype

159
00:07:30,840 --> 00:07:33,840
and ask about usability aspects.

160
00:07:33,840 --> 00:07:37,140
Also you have to stay open
minded and get excited

161
00:07:37,140 --> 00:07:39,450
about hearing hard truth.

162
00:07:39,450 --> 00:07:44,400
Don't try to hide or to
avoid any negative feedback.

163
00:07:44,400 --> 00:07:46,950
If you're really invested
in your product, project,

164
00:07:46,950 --> 00:07:51,420
or service, it's very easy
to overlook negative feedback

165
00:07:51,420 --> 00:07:55,080
and start selling your
product to respondents

166
00:07:55,080 --> 00:07:58,080
or cause them to be
unnecessary polite with you.

167
00:07:58,080 --> 00:08:01,500
So encourage and embrace negative feedback

168
00:08:01,500 --> 00:08:04,740
because it'll save you
effort going forward.

169
00:08:04,740 --> 00:08:09,740
Remember we said in lean
startup, fail early and often.

170
00:08:10,500 --> 00:08:12,210
We want to prevent the mistake

171
00:08:12,210 --> 00:08:15,960
so negative feedback is the most helpful.

172
00:08:15,960 --> 00:08:18,510
Active listening is also the key.

173
00:08:18,510 --> 00:08:22,560
It is important to help customers

174
00:08:22,560 --> 00:08:26,280
to provide their unbiased feedback to you.

175
00:08:26,280 --> 00:08:29,670
Open ended questions are most effective

176
00:08:29,670 --> 00:08:32,430
and definitely
significantly more effective

177
00:08:32,430 --> 00:08:34,080
than yes-no questions.

178
00:08:34,080 --> 00:08:39,080
Try to avoid binary answers
by opening your questions.

179
00:08:39,390 --> 00:08:42,660
So if a customer pauses,
do not try to fill the gap.

180
00:08:42,660 --> 00:08:44,340
Give them time to think.

181
00:08:44,340 --> 00:08:48,213
The goal is for you to learn,
not to market your product.

182
00:08:49,440 --> 00:08:53,310
And finally, we need to
aggregate and interpret the data.

183
00:08:53,310 --> 00:08:55,230
This step is extremely important

184
00:08:55,230 --> 00:09:00,230
because of the way the data
is aggregated and presented.

185
00:09:00,300 --> 00:09:03,360
This way will influence
your final decision.

186
00:09:03,360 --> 00:09:05,760
This is one of the reasons why the metrics

187
00:09:05,760 --> 00:09:09,960
and success criteria
are defined in advance.

188
00:09:09,960 --> 00:09:11,010
Based on the result,

189
00:09:11,010 --> 00:09:13,440
the decision is made whether to pivot

190
00:09:13,440 --> 00:09:14,850
or change the business model

191
00:09:14,850 --> 00:09:17,730
or one of the most
fundamental assumptions,

192
00:09:17,730 --> 00:09:21,030
or persevere and proceed
as originally intended.

193
00:09:21,030 --> 00:09:23,700
It's important to base decisions on data

194
00:09:23,700 --> 00:09:27,900
rather than speculations and
your interpretation of answers.

195
00:09:27,900 --> 00:09:30,513
Even if this is coming from the customers.

196
00:09:31,440 --> 00:09:36,060
As both Henry Ford and
Steve Jobs have shown,

197
00:09:36,060 --> 00:09:38,670
the features that customers ask for

198
00:09:38,670 --> 00:09:43,173
are never as interesting as
why they really want them.