recording-subtitles/2022/Intranet Analytics/output.srt

1
00:00:00,000 --> 00:00:13,120
Hello everyone, thank you for joining another session of MatomoCamp.

2
00:00:13,120 --> 00:00:21,320
Today we will learn more on how to get started with intranet analytics in Matomo.

3
00:00:21,320 --> 00:00:28,360
The host for this session is Marcus, who has been working with web since 1998 and with

4
00:00:28,360 --> 00:00:36,120
Matomo for the last 12 years. His experience from a public sector includes roles such as

5
00:00:36,120 --> 00:00:44,320
lead analyst, senior advisor in GDPR projects and product owner for Matomo and other analytics

6
00:00:44,320 --> 00:00:53,240
products. This session is recorded but Marcus will be available in the chat room and you

7
00:00:53,240 --> 00:00:58,960
can drop any questions there.

8
00:00:58,960 --> 00:01:07,400
Hi, let's begin the session on how to get started with intranet analytics in Matomo.

9
00:01:07,400 --> 00:01:12,960
This is the context, the things I will be talking about in this session. I think some

10
00:01:12,960 --> 00:01:19,040
of you would agree with me that web analytics is a common bad conscience. In my experience

11
00:01:19,040 --> 00:01:25,520
intranet analytics is even worse. I will be dealing with a couple of questions that is

12
00:01:25,520 --> 00:01:32,320
good to know to do intranet analytics right. For instance, how to regard consent as the

13
00:01:32,320 --> 00:01:40,000
legal basis to process employees' personal data. Why is Matomo a good choice for intranet

14
00:01:40,000 --> 00:01:46,280
analytics? And then I will guide you through the process of analytics, giving suggestions

15
00:01:46,280 --> 00:01:52,600
on how to make good use of Matomo. I will begin with introducing myself. I have been

16
00:01:52,600 --> 00:01:58,640
working in the industry since 1998 in different roles. I have been a consultant a couple of

17
00:01:58,640 --> 00:02:05,000
times and jumped the fence to work on the customer end as well in the Swedish public

18
00:02:05,000 --> 00:02:11,000
sector mostly. I contributed to get rid of Google analytics in favor of Matomo or back

19
00:02:11,000 --> 00:02:18,320
then peak before it was cool. In 2015, at Sweden's largest organization in the public

20
00:02:18,320 --> 00:02:24,840
sector, I have been responsible for search analytics, product owner for Matomo and performing

21
00:02:24,840 --> 00:02:31,360
a couple of advisor assignments such as implementing GDPR compliance and also curiously enough

22
00:02:31,360 --> 00:02:36,960
artificial intelligence in life science. In my spare time, I have been writing a couple

23
00:02:36,960 --> 00:02:41,680
of books on web strategy and web analytics. The latter is used as literature at some of

24
00:02:41,680 --> 00:02:47,640
Sweden's universities. Enough about me. Let's talk intranets. So what is the definition

25
00:02:47,640 --> 00:02:53,080
of an intranet in contrast to an extranet or a public website on the internet? Yeah,

26
00:02:53,080 --> 00:02:58,920
this might be both quite tricky and simple, right? Some people include everything connected

27
00:02:58,920 --> 00:03:04,880
in a form or another in the term internet. Some don't. I am one of those inclusive people.

28
00:03:04,880 --> 00:03:11,680
In my world, the internet includes internet, extranets and etc. But when using internet

29
00:03:11,680 --> 00:03:18,880
as a term, I think its main attribute is that it's an isolated network. Its users are the

30
00:03:18,880 --> 00:03:23,920
organization's own employees. And ideally, the internet facilitates the work to be done

31
00:03:23,920 --> 00:03:29,600
and making administrative tasks such as leave applications less of a burden compared to

32
00:03:29,600 --> 00:03:35,680
its paper equivalent. Regarding an internet's attribute as being inaccessible to non-employees,

33
00:03:36,240 --> 00:03:42,720
in Sweden, we also have a movement towards open internet when it comes to public sector

34
00:03:42,720 --> 00:03:48,240
organizations. Making the definitions less logical sometimes and legal compliance for

35
00:03:48,240 --> 00:03:54,960
analytics is not that straightforward. So what makes Matomo especially good at internet analytics?

36
00:03:54,960 --> 00:04:00,800
I possibly could ramble over the feats of using Matomo for way too long. But that shows two main

37
00:04:00,800 --> 00:04:08,080
points. First, to keep internal matters internal. With Matomo, you could actually gather personal

38
00:04:08,080 --> 00:04:12,640
data that would be breach of contract with some of the other popular products on the market.

39
00:04:13,280 --> 00:04:19,040
My second point is the more legal aspects of compliance, as in not sharing data with

40
00:04:19,040 --> 00:04:27,120
unnecessary outsiders, especially those with ties to third country legislation, in essence,

41
00:04:27,120 --> 00:04:33,200
United States. If you do, then you have exposed the data to actors you and your employees cannot

42
00:04:33,200 --> 00:04:39,200
hold accountable on what happens with that data. It's irresponsible. Also, to gather

43
00:04:42,400 --> 00:04:47,360
data on internet use has its own unique legal problems, as I will address shortly.

44
00:04:47,360 --> 00:04:53,760
Thirdly, for a smaller organization, it's not that hard to get going. Even without an IT department,

45
00:04:53,760 --> 00:04:58,880
you could make use of a web host, since many of them have one-click installers for Matomo.

46
00:04:59,920 --> 00:05:06,560
Now I am going to dive into the details. There might be differences in terms of configuration

47
00:05:06,560 --> 00:05:12,000
between Matomo internet tracking and Matomo for a public website. Technically, if you have

48
00:05:12,000 --> 00:05:16,480
another Matomo setup for your internet, you don't really have to worry about the same type of

49
00:05:16,480 --> 00:05:22,320
things as with a Matomo reachable from the public internet. The internet is not exposed in the same

50
00:05:22,320 --> 00:05:28,400
way. On an internet property in Matomo, you get a more even load, for instance, since the things

51
00:05:28,400 --> 00:05:35,120
being tracked is behind a firewall, it is not exposed to refer, spam, or denial of service attacks.

52
00:05:36,000 --> 00:05:42,560
Perhaps the most important point is to own your data, being able to use and reuse the data when

53
00:05:42,560 --> 00:05:50,800
you see fit. A Swedish term that might be lost in translation is rådighet. It translates to

54
00:05:50,800 --> 00:05:57,280
resourcefulness, which is not bad, but it lacks, in my opinion, the comforting and trustworthiness

55
00:05:57,280 --> 00:06:04,160
in its Swedish counterpart. Matomo as a product gives you control and sovereignty if you want to.

56
00:06:04,880 --> 00:06:10,160
By the way, I am not advocating solely for self-hosting Matomo, but when you don't have

57
00:06:10,160 --> 00:06:16,160
to make an effort to look at the processor and the eventual sub-processor of the data in question.

58
00:06:16,800 --> 00:06:22,880
What are the rules in terms of privacy when it comes to an internet? The e-privacy regulation in

59
00:06:23,680 --> 00:06:30,640
the European Union is not automatically applicable for internets, but if the user is not an employee,

60
00:06:30,640 --> 00:06:36,240
then yes, because sometimes the user of an internet is a journalist, a member of the public,

61
00:06:36,240 --> 00:06:42,080
or a non-employee. If you have never heard of European e-privacy regulation, you might have

62
00:06:42,080 --> 00:06:49,120
heard of something on the lines of the cookie law. It was introduced in the early 2000s and recently

63
00:06:49,120 --> 00:06:57,200
updated. In an EU perspective, all of us got the GDPR decade and the ifs, hows and whens of processing

64
00:06:57,200 --> 00:07:03,440
personal data, and all form of handling of personal data is considered processing, such as the

65
00:07:03,440 --> 00:07:10,320
collection, processing, sharing, etc. The only way we can process personal data is if we have at

66
00:07:10,320 --> 00:07:19,360
least one of six of the conditions that apply. That is A to F in this slide. All other cases are not

67
00:07:19,360 --> 00:07:26,800
lawful. I worked as a web analytics specialist in Sweden's largest region when we had the GDPR

68
00:07:26,800 --> 00:07:33,520
project, and it was really obvious during talks with all our lawyers that consent is not great on

69
00:07:33,520 --> 00:07:39,680
a public website, and consent is even less applicable on an internet, since the relation

70
00:07:39,680 --> 00:07:46,480
between the employee and employer is not living up to the voluntary part of the consent requirements

71
00:07:46,480 --> 00:07:54,480
stated in the GDPR. Consent is not lawful if it's not freely given, as in there is an option to turn

72
00:07:54,480 --> 00:08:01,680
it down. So that dictates if we can ask for consent on an internet, and the other law, the e-privacy

73
00:08:01,680 --> 00:08:08,720
regulation, makes asking for cookies look like a consent that might itself be illegal on the GDPR,

74
00:08:08,720 --> 00:08:16,720
if on an internet, or employees at least. The way to process employees' data, such as IP addresses,

75
00:08:16,720 --> 00:08:22,960
for an internet analytics, could possibly be lawful through B, the contract between the employees

76
00:08:22,960 --> 00:08:29,200
and the employer. This is rather difficult, but worth having in mind when an organization makes

77
00:08:29,200 --> 00:08:35,360
up their mind on the Matomo implementation for their internet. The easy way out is to use log

78
00:08:35,360 --> 00:08:41,760
analytics, to at least not having to consider e-privacy. But data in logs still is regulated

79
00:08:41,760 --> 00:08:48,320
by GDPR, and the default is that any personal data only belongs to the data subject, the person,

80
00:08:48,320 --> 00:08:54,160
in question. So how to set up an internet analytics with Matomo by yourself, in order to have a

81
00:08:54,160 --> 00:09:00,960
sandbox to test Matomo for your internet. I am myself playing with Matomo at a local webhost

82
00:09:00,960 --> 00:09:08,240
when stakes are low. Think of it as playing in a sandbox. For anyone interested in a one-click

83
00:09:08,240 --> 00:09:14,720
installer on Matomo, when having a webhost, the package is present in Softaculous, for instance.

84
00:09:14,720 --> 00:09:21,520
This is of course not suitable for everyone. Well, the time when major design decisions were made

85
00:09:21,520 --> 00:09:27,680
based on subjective measures, like someone's personal taste, ought to be behind us. Now you

86
00:09:27,680 --> 00:09:33,680
can and should evaluate almost everything based on data. Previously performed tests and the

87
00:09:33,680 --> 00:09:38,960
insights gained from working with analytics could be readily available to internet teams. Instead of

88
00:09:38,960 --> 00:09:45,120
attending pointless meetings about what shade of blue you like, you can evaluate alternatives based

89
00:09:45,120 --> 00:09:51,520
on real user of your internet. Users, through their clicks, provide signals on what works,

90
00:09:51,520 --> 00:09:57,200
how to find things, and which routes they take. This talk is about how to get started with the

91
00:09:57,200 --> 00:10:03,840
analytics in an initial review of an internet. Often no one knows the usefulness of most of the

92
00:10:03,840 --> 00:10:09,280
internet content, and there is little effort toward measuring the impact of the internet

93
00:10:09,280 --> 00:10:16,720
and the performance of the business, at least beyond noting the number of users. For many teams,

94
00:10:16,720 --> 00:10:22,400
the reason is they do not have a training in basic analytics, and which all the statistics

95
00:10:22,400 --> 00:10:29,280
to make use of. I hope you want to start evaluating your internet in a structured way and set tangible

96
00:10:29,280 --> 00:10:34,960
business goals for measuring what success looks like on the internet. The internet is not just

97
00:10:34,960 --> 00:10:42,000
a collection of pages, images, and documents. Rather, it's there to support employees in their

98
00:10:42,000 --> 00:10:48,800
daily work, efficiently carrying out work tasks like reserving a place on the training course,

99
00:10:49,600 --> 00:10:55,680
filing an application for leave, or finding the phone number of the person in the human resource

100
00:10:55,680 --> 00:11:03,280
department. Working with web analytics for an internet is about improving or simplifying

101
00:11:03,280 --> 00:11:09,680
activities like these, activities that the internet should support. Internet analytics is not

102
00:11:09,680 --> 00:11:15,520
about accumulating vast amounts of data. It is about using the data to gain insights in the

103
00:11:15,520 --> 00:11:21,040
user's experience of the internet, with the intent of improving that experience.

104
00:11:21,040 --> 00:11:28,400
It is easy to focus on analyzing the collected traffic data. However, all kinds of tools can

105
00:11:28,400 --> 00:11:33,680
contribute to getting an overview of your internet performance, and should be included in the concept

106
00:11:33,680 --> 00:11:40,400
of internet analytics. Some of these tools are soulless premium plugins for Matomo, such as

107
00:11:40,400 --> 00:11:48,080
session recording, but Matomo also evolves. I think it will be in version 4.1, which will be

108
00:11:48,080 --> 00:11:56,000
evolved. I think it will be in version 4.0. The performance report will be introduced out of the

109
00:11:56,000 --> 00:12:03,120
box. Problems such as not fulfilling accessibility requirements, using unnecessarily high resolution

110
00:12:03,120 --> 00:12:09,600
images, and many more may be easy to identify without using statistics. But with Matomo,

111
00:12:09,600 --> 00:12:14,960
you could develop a plugin to enhance the dashboard with that kind of data as well.

112
00:12:14,960 --> 00:12:20,240
A key performance indicator, KPI, sometimes, sometimes for internet schemes, we call it the

113
00:12:20,240 --> 00:12:26,000
metric. It's a measurable value that demonstrates how effectively business objectives are being

114
00:12:26,000 --> 00:12:34,320
attained. Organizations use KPIs to evaluate their success as reaching their targets. There are two

115
00:12:34,320 --> 00:12:40,320
fundamental categories of KPIs and metrics. We need them both to understand the big picture.

116
00:12:40,320 --> 00:12:47,840
Quantitative measurements are figures on how the internet is performing, such as all the data

117
00:12:47,840 --> 00:12:54,000
about the users as a group we find in Matomo, for instance. These measurements provide answers on

118
00:12:54,000 --> 00:13:02,240
how, when, and what about users, what devices they use when they visit the internet, if they

119
00:13:02,240 --> 00:13:07,520
have actively interacted with internet content in any way, where did they come from, and so on.

120
00:13:07,520 --> 00:13:15,920
Simply put, behavioral data. Qualitative measures is about how users experience the internet. This

121
00:13:15,920 --> 00:13:23,280
data is in the form of text or film rather than numbers. The methods may include an employee

122
00:13:23,280 --> 00:13:30,480
survey, guerrilla testing, interviews, usability evaluations, user observation, or features on the

123
00:13:30,480 --> 00:13:36,960
site where the user can report what they thought of their experience. In some cases, we hope to

124
00:13:36,960 --> 00:13:43,680
serve the user so we can take the opportunity to ask follow-up questions if necessary. Your internet

125
00:13:43,680 --> 00:13:50,880
should have a regular ongoing review process and major review should take place if you are considering

126
00:13:50,880 --> 00:13:58,880
an upgrade or change of platform. Reviews are often called content audits, although their

127
00:13:58,880 --> 00:14:05,280
conclusions are not restricted to the quality of content. When reviewing the current internet, both

128
00:14:05,280 --> 00:14:11,920
quantitative and qualitative measures can be used to find out what is working and what needs to be

129
00:14:11,920 --> 00:14:19,040
improved. The basic level of review should enable you to group content into useful pages that have

130
00:14:19,040 --> 00:14:26,320
an obvious value to actual users and all the rest at the best dubious value. When looking through the

131
00:14:26,320 --> 00:14:31,600
content, there are some things to consider when deciding which groups the content belongs to,

132
00:14:31,600 --> 00:14:37,520
which parts of your internet support the important activities. This part needs to be analyzed to

133
00:14:37,520 --> 00:14:43,360
identify improvement opportunities, optimal placement of buttons, understandable texts,

134
00:14:43,360 --> 00:14:50,240
how easy it is to get there, etc. How easy it is to find your way between the different activities.

135
00:14:51,600 --> 00:14:57,360
Then consider your navigation structure. Is it easy to find your way back to a starting point if

136
00:14:57,360 --> 00:15:04,880
you got a direct link by email? Is it easy to switch between tasks? Is it easy to switch from

137
00:15:04,880 --> 00:15:12,160
applying for vacation to the task to correct an improperly filed holiday application?

138
00:15:13,680 --> 00:15:20,960
How easy is it for users to complete an activity? Just like a store survival is conditional on

139
00:15:20,960 --> 00:15:27,760
customers' ability to pay for the goods, you need to make sure there are as few obstacles as possible

140
00:15:27,760 --> 00:15:35,920
on your internet in order for users to complete an activity. What defines a successful session on

141
00:15:35,920 --> 00:15:41,680
the internet? What activities must have occurred for you to be able to objectively refer to a

142
00:15:41,680 --> 00:15:48,480
user's session on the internet as a successful visit? Is it enough that the user fills out their

143
00:15:48,480 --> 00:15:54,320
timesheet or maybe that they land on home page, read some internal news and then disappear?

144
00:15:55,200 --> 00:16:01,840
Think carefully about which behavior indicates a successful visit. Content that does not relate to

145
00:16:01,840 --> 00:16:07,120
this point is often distracting and it should probably be removed or at least be hidden out

146
00:16:07,120 --> 00:16:13,280
of sight and not indexed by your search engine. Please do define requirements for new pages to

147
00:16:13,280 --> 00:16:20,560
include a clearly displayed call to action and having an explicit connection to the goals for

148
00:16:20,560 --> 00:16:30,320
the internet. Talking about call to actions, I just mentioned it. This is the definition of the term.

149
00:16:30,320 --> 00:16:38,000
Often a CTA is a button or any other interactive design element that is enabling the user to take

150
00:16:38,000 --> 00:16:45,280
the next logical step towards a conversion on a goal. This process is what the rest of my talk

151
00:16:45,280 --> 00:16:51,840
will discuss. For analytics to be meaningful, you need to first define what is good. Good content,

152
00:16:51,840 --> 00:16:57,840
successful task accomplishment, good quality and what is good for business. These are the things

153
00:16:57,840 --> 00:17:04,800
that are worth analyzing. Otherwise, the risk is that you are working to improve things that lack

154
00:17:04,800 --> 00:17:11,280
the prospect of worthwhile improvement. The working process for internet analytics is as follows.

155
00:17:12,080 --> 00:17:18,240
You start at the upper left of this never-ending loop. First, develop business goals or evaluate

156
00:17:18,240 --> 00:17:25,440
existing business goals. Then you produce reports and develop methods for analyzing

157
00:17:25,440 --> 00:17:35,520
KPIs and selecting metrics. Thirdly, you analyze and fourthly improve the internet based on any

158
00:17:35,520 --> 00:17:42,480
findings in the analysis and then you iterate. This is not the project. Start by developing the

159
00:17:42,480 --> 00:17:48,400
business goal and you're never done. Begin the second iteration with reflection of the measurability

160
00:17:48,400 --> 00:17:55,200
of the business goals that can be improved on another iteration of the process. Now it's time

161
00:17:55,200 --> 00:18:00,320
to go through each step in this working process and examine what is worth thinking about.

162
00:18:01,840 --> 00:18:10,240
Last year, it's pretty mind-boggling that only 34% of internet managers, according to

163
00:18:10,240 --> 00:18:18,480
web service awards, still had no clear objective for their internet. A clear majority also say

164
00:18:19,120 --> 00:18:23,040
that they have no defined objectives for any of the internet's sub-pages.

165
00:18:23,040 --> 00:18:27,760
Obviously, it's difficult to know how useful the internet is if you do not know what it's supposed

166
00:18:27,760 --> 00:18:33,840
to achieve or how to measure it. The very first step of internet analytics is to set down a list

167
00:18:33,840 --> 00:18:43,200
of goals. You should avoid vaguely worded wishes such as enable findability or keep employees

168
00:18:43,200 --> 00:18:50,320
informed with news. Instead, try to stick with concrete activities and tasks employees need to

169
00:18:50,320 --> 00:18:56,320
be able to do to get their job done and bring value to the business. Good examples of these include

170
00:18:56,960 --> 00:19:02,800
reporting time at work, updating skills profile, or submitting requests for absence.

171
00:19:04,240 --> 00:19:12,960
Management documents such as the organization's vision, plans, and the budget often provide

172
00:19:12,960 --> 00:19:20,320
already defined objectives, KPIs, that can be reused or adapted as measures for your internet.

173
00:19:20,320 --> 00:19:25,120
Break down each goal to something that can be isolated and measured as part of the user

174
00:19:25,120 --> 00:19:31,760
experience of the internet. A challenge may be to align the measurable business objectives with

175
00:19:31,760 --> 00:19:37,520
the long-term objectives. It is important to try to anticipate the long-term consequences of

176
00:19:37,520 --> 00:19:44,000
the goals we define in order to avoid negative side effects. In addition to the business objectives,

177
00:19:44,000 --> 00:19:50,880
it can be beneficial to review other metrics such as the number of page views per visit.

178
00:19:51,440 --> 00:19:58,080
They may tell the backstory of why the objectives performed the way they did. However,

179
00:19:58,800 --> 00:20:05,920
most generic metrics struggle to prove what value they indicate is the result unequivocally good

180
00:20:05,920 --> 00:20:14,240
or bad. These are not where you should put most of your effort. Instead, they help you investigate

181
00:20:14,240 --> 00:20:21,600
things afterwards and gather data helping you to tell the story of the users when you are reporting.

182
00:20:22,800 --> 00:20:29,520
The amount of page views, unique users, sessions, and similar does not provide any evidence of

183
00:20:29,520 --> 00:20:35,680
increased business performance. These may be nice anecdotes between internet geeks, but they provide

184
00:20:35,680 --> 00:20:41,680
very little insight in which way the internet is successful. Instead, they make for personal

185
00:20:41,680 --> 00:20:48,960
glorifications and a good gut feeling of not great business. Another term is segmentation. It is a

186
00:20:48,960 --> 00:20:53,840
feature of Matuomo and you find it at the toolbar in the top of the user interface.

187
00:20:54,640 --> 00:21:01,040
A segmentation is a group or a subset of users that have some attribute in common. It may be

188
00:21:01,040 --> 00:21:07,760
that they all use a specific type of device. It can be the geographical location or whatever Matuomo

189
00:21:07,760 --> 00:21:13,920
has gathered as common data points for them. Segments are often overlooked, so I will take

190
00:21:13,920 --> 00:21:22,400
some time to pitch it to you. But before I go into even more segmentation, we have a fundamental

191
00:21:22,400 --> 00:21:30,560
term to discuss, click-through rate or CTR for short. CTR is the share of users that do a specific

192
00:21:30,560 --> 00:21:37,840
thing, such as that, for instance, 73% of users click on one of the featured articles on the front

193
00:21:37,840 --> 00:21:45,680
page. If you have a low CTR, that indicates the need to improve the usability of the place

194
00:21:46,320 --> 00:21:53,360
where many users disappear. Or perhaps they get cold feet because of some poor design decision

195
00:21:53,360 --> 00:21:59,760
you need to improve. Think about the goal that at least 90% should succeed with submitting their

196
00:21:59,760 --> 00:22:06,160
application of lead. If 25% of them drop out at the very first step in the process, then we have

197
00:22:06,160 --> 00:22:11,680
to try to figure out what caused it or possibly set a more reasonable goal we are actually able

198
00:22:11,680 --> 00:22:19,200
to achieve. It is particularly interesting to create a segment of users to find out when we

199
00:22:19,200 --> 00:22:25,280
are failing the users. The segment, those who do not complete all the steps in the conversion funnel,

200
00:22:25,280 --> 00:22:32,480
might reveal some interesting pattern. Sometimes you are able to alleviate. Are there any common

201
00:22:32,480 --> 00:22:40,160
denominators where the users fail or flee? Where do they go? Is there possibly an indistinct call

202
00:22:40,160 --> 00:22:47,440
to action that users do not understand, see or discover? The conversion funnel does not require

203
00:22:47,440 --> 00:22:54,160
many steps to be suitable as a visualization. It might be obvious in itself presenting the results

204
00:22:54,160 --> 00:23:00,320
to internal stakeholders. A simple use case for a conversion funnel can be to show the amount of

205
00:23:00,320 --> 00:23:06,320
users of the search engine that actually click on any of the results. How many times they rephrase

206
00:23:06,320 --> 00:23:13,200
the search terms and how big the drop-off is. It could be one of several means to evaluate whether

207
00:23:13,200 --> 00:23:19,440
an enhanced relevance model for the search engine has been an improvement according to its users.

208
00:23:19,440 --> 00:23:25,280
A conversion funnel is not limited to the internet. It could very well be the measure

209
00:23:25,280 --> 00:23:32,080
of impact of a newsletter by email or even something in the real world. One thing is certain

210
00:23:32,080 --> 00:23:38,640
though. Do not report data or specific numbers, so-called data puking. It ought to be the last

211
00:23:38,640 --> 00:23:45,760
option to try out on stakeholders. Very few are interested in data alone. Even fewer actually

212
00:23:45,760 --> 00:23:52,400
understand the data. You are supposed to tell them the story of the data and visualize the impact.

213
00:23:52,400 --> 00:23:59,600
Conversion funnels are in a way the opposite of data puking since it's in a visual way telling

214
00:23:59,600 --> 00:24:08,160
a story reminding you to explain where the users end up and your hypothesis on why. Back to segments.

215
00:24:08,160 --> 00:24:13,200
Here I'll make some suggestions on segments in Matomo for internets for you to think about.

216
00:24:13,200 --> 00:24:18,480
A feature of Matomo that, at least in my experience, many seem to overlook is to compare two

217
00:24:18,480 --> 00:24:25,760
segments against each other. My first suggestion is such a comparison. Depending on the content of

218
00:24:25,760 --> 00:24:30,880
your internet, there might be a use case for a segment singling out users with a management

219
00:24:30,880 --> 00:24:36,800
position. This is something you can inform Matomo about using the internet as an example.

220
00:24:36,800 --> 00:24:42,320
You can inform Matomo about using the custom dimension feature directly from the code of your

221
00:24:42,320 --> 00:24:49,440
internet. So, what is the point? Well, if you have content on the internet aiming towards users

222
00:24:49,440 --> 00:24:55,120
within management, the obvious question is if that content is used by its intended audience, right?

223
00:24:56,480 --> 00:25:03,040
This is how you get to figure that out. Other suggestions on segments is the user's organizational

224
00:25:03,040 --> 00:25:09,760
position, since it can be relevant as a context when looking at conversions on the internet.

225
00:25:09,760 --> 00:25:14,960
For instance, if a part of the organization is unexpectedly underperforming on some of the

226
00:25:14,960 --> 00:25:22,000
internet goals, you have a deviation to investigate further. Another suggestion is to segment on

227
00:25:22,000 --> 00:25:28,080
different geographic locations or perhaps different office buildings in the same city or floors in the

228
00:25:28,080 --> 00:25:36,800
same building. So, how exactly is custom segments done in Matomo? One way is to let the internet

229
00:25:36,800 --> 00:25:42,800
feed the info to Matomo through the front-end code to end up in a custom dimension. An internet

230
00:25:42,800 --> 00:25:49,200
usually has integration to the local directory of employees, who is logged on, the actual computer,

231
00:25:49,200 --> 00:25:55,440
and where do they work in the organization. That might make the cut for a custom dimension,

232
00:25:55,440 --> 00:26:03,760
since they ought not to be of temporary nature or triviality for a few involved. But there might

233
00:26:03,760 --> 00:26:09,920
also be possibilities to track signals for segments using Matomo's event tracking. The upside with

234
00:26:09,920 --> 00:26:17,040
the Matomo's way of having segments as inclusive filters is that if a user session consists of an

235
00:26:17,040 --> 00:26:25,600
event signal of your choice, they are segment compatible. And you can see this slide as a

236
00:26:25,600 --> 00:26:32,560
critique on the sometimes oversimplistic views on segments, personas, and other ways of describing

237
00:26:32,560 --> 00:26:39,520
our users. Another difficulty is which potentially crucial details are hidden behind the persona or

238
00:26:39,520 --> 00:26:46,400
target group we have. Here we have an illustrative example some of you may have seen previously.

239
00:26:46,400 --> 00:26:52,080
It is the criticism of personas as a concept. That is, that you create a fictional person who

240
00:26:52,080 --> 00:27:00,240
is named and assigned a number of characteristics. But criticism is useful in many contexts.

241
00:27:01,440 --> 00:27:08,240
I have put together a persona I call Welfare William. He is rich, has children, is British,

242
00:27:08,240 --> 00:27:17,520
homeowner, etc. The description fits both the, until recently, Prince Charles and Ossie Osborn,

243
00:27:17,520 --> 00:27:25,040
two quite different people. One of them lived 70 years of his life as a literal prince. The other

244
00:27:25,040 --> 00:27:32,720
one is a prince of sorts, the prince of darkness, perhaps. One had fried sparrows flying into his

245
00:27:32,720 --> 00:27:39,680
mouth since birth. The other has literally bitten the necks of bats on concept stages, so the blood

246
00:27:39,680 --> 00:27:46,080
spurted. One is the very definition of upper class, the other began his life in total social

247
00:27:46,080 --> 00:27:53,680
vulnerability. What unites Ossie and the now King Charles may actually be irrelevant, but they have

248
00:27:53,680 --> 00:28:00,800
very different life experiences. And if you were to design a communication campaign and are aware

249
00:28:00,800 --> 00:28:07,520
that these two groups exist within your target group, yes, I suppose you will see the challenge

250
00:28:07,520 --> 00:28:14,080
in how the appeal is, how flexible these two people are to act on your call to action.

251
00:28:15,760 --> 00:28:21,680
There are two things you can take away from this discussion, though. One is that personas should

252
00:28:21,680 --> 00:28:29,120
not be about demographic data, such as income, marital status, etc. Rather, it is the person's

253
00:28:29,120 --> 00:28:36,320
problem to solve and their challenges, because that's where you become relevant

254
00:28:36,320 --> 00:28:43,440
with whatever matching offer or service you have. The second one is to dig deeper, otherwise you

255
00:28:43,440 --> 00:28:51,120
may miss details that are possibly relevant. So I leave it up to you and the internet managers to

256
00:28:51,120 --> 00:28:59,200
figure out if management position is a relevant segment where you work, or if it's just another

257
00:28:59,200 --> 00:29:09,200
wealthy William case. Segments help you not to get lost in all the averages and aggregates you

258
00:29:09,200 --> 00:29:17,360
encounter in Matomo. A made-up illustration is now in this image, showing that the fact that the

259
00:29:17,360 --> 00:29:23,680
average user, the grey circle in the middle, has 10 page views per visit might not be that insightful

260
00:29:23,680 --> 00:29:31,680
if actual users typically split into two groups with either one page view or 20. Segmentation is

261
00:29:31,680 --> 00:29:37,760
important because of the risks of looking at average values. Average values are sometimes not

262
00:29:37,760 --> 00:29:44,160
representative for the majority of users, and average value is meaningful only if the data is

263
00:29:44,160 --> 00:29:51,120
in a normal distribution, meaning that the average value has relatively few surprising deviations.

264
00:29:53,360 --> 00:30:00,320
This is an illustration of a normal distribution. Most data is predictably similar to the average

265
00:30:00,320 --> 00:30:07,840
value. Not all things we are interested in is distributed this way, just as the made-up example

266
00:30:07,840 --> 00:30:16,240
of average value of 10 page views per visit is not that meaningful a value. Sometimes the average

267
00:30:16,240 --> 00:30:22,480
does not describe the reality. More interesting is the observation that there are at least two

268
00:30:22,480 --> 00:30:28,080
very different groups in this data. To get clarity around what groups you need to inspect,

269
00:30:28,080 --> 00:30:34,320
you use segmentation to divide the data into different groups that can be explored separately.

270
00:30:34,320 --> 00:30:40,720
The idea is to filter out a subset of the users, see how they perform, and if they stand out in

271
00:30:40,720 --> 00:30:48,720
some way. Segmentation is done to be aware of differences, differences in how each segment

272
00:30:48,720 --> 00:30:54,320
behaves, depending on where they are located, their profession, which sub-pages they visit,

273
00:30:54,320 --> 00:31:00,960
and so forth. When reviewing your internet using analytics, it is a good idea to focus on

274
00:31:00,960 --> 00:31:08,800
to focus on few segments per repetition of the analysis. One segment could be to filter out

275
00:31:08,800 --> 00:31:15,120
users who only have a single page view and inspect where those user sessions are taking place.

276
00:31:16,240 --> 00:31:21,680
Another segment could be to watch the behavior among those who use the internet on a daily basis,

277
00:31:22,320 --> 00:31:29,200
what differentiates them from monthly users, for instance. Yet another segment may be to look at

278
00:31:29,200 --> 00:31:35,840
those who repeatedly only make a single page view. Looking at these different usage patterns

279
00:31:35,840 --> 00:31:41,280
could lead to some significant improvement actions. If, for example, you conclude that

280
00:31:41,280 --> 00:31:48,080
from your analysis that the single page viewer segment have simply given up on the internet,

281
00:31:48,080 --> 00:31:53,920
the targeted campaign could be implemented, aiming at figuring out a way of convincing them

282
00:31:53,920 --> 00:32:00,720
to try out some new features. As mentioned previously, average values do not always provide

283
00:32:00,720 --> 00:32:07,840
a particularly accurate picture of reality or actionable insights. The majority of users may have

284
00:32:07,840 --> 00:32:14,240
few usability issues. The aim of segmentation is to reveal the portion of users within the whole

285
00:32:14,240 --> 00:32:21,120
who have problems. With internet analytics, you can simultaneously manage several user groups

286
00:32:21,120 --> 00:32:29,120
in a structured way. Examples of business and operational goals measurable on an internet.

287
00:32:29,120 --> 00:32:37,440
Content published on the internet needs to align with business goals and the intentions we have

288
00:32:37,440 --> 00:32:44,000
for the internet. It is not nearly enough to assert a public interest or to continue treating

289
00:32:44,000 --> 00:32:49,600
the internet as an information dump just because some of us like to produce content.

290
00:32:49,600 --> 00:32:57,680
I have set out some examples of business goals for the internet, such as that at least 90% of

291
00:32:57,680 --> 00:33:05,360
those who begin the process to apply for BLEEM through the internet have to succeed. Otherwise,

292
00:33:05,360 --> 00:33:17,600
they are wasting their time. Perhaps you could have efficiency metrics as well.

293
00:33:17,600 --> 00:33:23,760
Metrics such as employees should easily find their contacts within the human resources

294
00:33:23,760 --> 00:33:30,320
or finance department whenever needed. That is very much a task to be done

295
00:33:31,440 --> 00:33:38,080
thing for the internet. As you notice, these examples of business goals are not limited by

296
00:33:38,080 --> 00:33:44,080
what statistics are captured automatically in Matuomo. To measure real business goals,

297
00:33:44,080 --> 00:33:52,960
you have to work both with quantitative metrics and qualitative metrics such as

298
00:33:52,960 --> 00:33:59,200
users' subjective opinions. Depending on the business goal, you might also need to look at

299
00:33:59,200 --> 00:34:06,880
other systems than Matuomo. Regarding accessibility and a measurable way to evaluate it,

300
00:34:06,880 --> 00:34:14,640
the design should probably adhere to accessibility guidelines. One such guideline is

301
00:34:16,160 --> 00:34:26,480
VCAG 2.2 at level AA, which in the European Union is a mandatory regulation since

302
00:34:27,360 --> 00:34:31,200
two years ago, I think, at least for the public sector.

303
00:34:31,200 --> 00:34:40,560
Step two out of four of the process, then. If you mention internet, you probably have access to some

304
00:34:40,560 --> 00:34:46,240
standard reporting tools. I would also guess that tool might be Matuomo since you attend this

305
00:34:46,240 --> 00:34:53,600
conference or that you are considering Matuomo in the future. In order to monitor how well the

306
00:34:53,600 --> 00:34:59,520
internet delivers on goals and metrics, we need to gather appropriate data using these data.

307
00:34:59,520 --> 00:35:06,240
It's almost certainly necessary to tune your reporting tool so that your

308
00:35:06,240 --> 00:35:13,040
reports actually measure your business goals. When you need adjustments or variations of your

309
00:35:13,040 --> 00:35:18,960
standard reports in Matuomo, you can consider premium plugin custom reports. In my experience,

310
00:35:18,960 --> 00:35:26,160
that plugin is a lifesaver. It can give you customized reports, filter data, and answer

311
00:35:26,160 --> 00:35:32,080
metric questions without the need to over and over again look at multiple other reports.

312
00:35:35,920 --> 00:35:41,840
It is good to collect data that tells the story about the goal so that we learn why something

313
00:35:41,840 --> 00:35:49,280
performed better, not only that it actually did. It may prove useful to work with a web developer

314
00:35:49,280 --> 00:35:54,480
until you have gained some knowledge of the systems involved. A large part of the benefit

315
00:35:54,480 --> 00:36:00,400
of working with the reports and methods is the learning process. Working with the system itself

316
00:36:00,400 --> 00:36:06,960
to get to know the user's behaviors and needs is something that makes each iteration of this work

317
00:36:06,960 --> 00:36:14,640
begin on a new and higher level. There are a number of widely used techniques, methods,

318
00:36:14,640 --> 00:36:20,640
and templates we can use to improve how your reports are designed and I will go through some

319
00:36:20,640 --> 00:36:31,440
of these now. First off is the conversion funnel, also premium plugin. This is funnel visualization.

320
00:36:32,560 --> 00:36:37,200
It shows that no one followed the desired path from the landing page to one at the top.

321
00:36:37,840 --> 00:36:46,560
This funnel has that first step as optional and we can see in the middle that users coming in

322
00:36:46,560 --> 00:36:54,400
the funnel on that level converts at 25 percent rate towards the goal called signed up. A funnel

323
00:36:54,400 --> 00:36:59,680
like this indicates room for improvement. Probably we are distracting the users with

324
00:36:59,680 --> 00:37:05,920
more tempting content or in this case the landing page is not optimized as pushing users down the

325
00:37:05,920 --> 00:37:13,440
funnel. The conversion is when a user action results in a measurable contribution towards

326
00:37:13,440 --> 00:37:21,360
a business goal, for instance signing up for the corporate newsletter or being successful in some

327
00:37:21,360 --> 00:37:28,640
self-service process at the internet. Probably the most common method to visualize a multi-step

328
00:37:28,640 --> 00:37:34,160
process is to use the conversion funnel. It is particularly well suited to visualize where

329
00:37:34,160 --> 00:37:41,040
users fail in a process with several stages. You define the starting point and then measure the

330
00:37:41,040 --> 00:37:47,840
click-through rate on each part, that is how many users you lose on the road towards the goal.

331
00:37:49,440 --> 00:37:54,320
All objectives cannot be measured by using the collected data, the already collected data in

332
00:37:54,320 --> 00:38:01,600
Matomo, but some do depending on your implementation. It is not solely about users click and their

333
00:38:01,600 --> 00:38:08,480
behavior. We need to evaluate other forms of quality indicators. This is a list of quality

334
00:38:08,480 --> 00:38:17,520
indicators not drawn from statistics tools that a communications department can use to measure

335
00:38:17,520 --> 00:38:28,240
quality. Things like to list heavy images that are slow to load especially from mobile users,

336
00:38:29,920 --> 00:38:31,360
find duplicate content,

337
00:38:31,360 --> 00:38:38,720
documents called something inappropriate such as new word document, has no title at all or other

338
00:38:38,720 --> 00:38:48,240
common mistakes that make it difficult to find the content on the internet, or that content lacks

339
00:38:48,240 --> 00:38:55,440
crucial metadata like page description or keywords to help them turn on search engine.

340
00:38:55,440 --> 00:39:01,200
Talking about search, an easy test you can perform on your own internet is to search for what's not

341
00:39:01,200 --> 00:39:08,320
supposed to give any results, as in the illustration here when I search for Microsoft Word which ends

342
00:39:08,320 --> 00:39:16,880
up being the document title if a user is foundling. Where I used to work until recently we have about

343
00:39:16,880 --> 00:39:24,000
one million pages and documents in our search index with the help of the search engine. You can

344
00:39:24,000 --> 00:39:29,360
behind the scenes compile a lot of interesting statistics, at least if you consider the search

345
00:39:29,360 --> 00:39:36,800
feature just as any other user interface to analyze. For instance we could see from search

346
00:39:36,800 --> 00:39:43,440
robot statistics that 89 percent of the internet lacks keywords, but we could also see that

347
00:39:43,440 --> 00:39:49,840
statistics that 89 percent of the internet lacks keywords, in other words nine out of ten

348
00:39:49,840 --> 00:39:55,600
pages and documents are not really competing in the battle to reach the top of the internal search

349
00:39:55,600 --> 00:40:02,000
engine. Then it's no wonder that the search engine has a hard time figuring out what is

350
00:40:02,000 --> 00:40:10,080
the most relevant content to some search queries. Most internets do have some sort of search feature,

351
00:40:10,080 --> 00:40:16,640
at the very least you ought to try to learn what keywords or phrases the users use. An extremely

352
00:40:16,640 --> 00:40:24,080
actionable thing is if you also track searches that return no result, zero hits. If a search phrase is

353
00:40:24,080 --> 00:40:32,880
reasonable then it should have a result. You might have issues with synonyms or that perhaps vital

354
00:40:32,880 --> 00:40:39,440
content is lacking. You could also consider to track the click behavior in the search results

355
00:40:39,440 --> 00:40:45,600
there is an expected distribution on how much clicks the first entry would get, the second

356
00:40:45,600 --> 00:40:53,360
would get about half of that etc. And perhaps you would track clicks that are not the top three in

357
00:40:53,360 --> 00:41:00,320
the results or that an entry got twice as much clicks than is expected for its ranking in the

358
00:41:00,320 --> 00:41:08,160
results etc etc. And I highly recommend Louise Rosenfeld's book called Search Analytics for

359
00:41:08,160 --> 00:41:14,880
Your Site if you are curious about a method to evaluate search features and ranking.

360
00:41:16,320 --> 00:41:21,680
The next step is of course to look at, compare and analyze all the data you have gathered.

361
00:41:22,480 --> 00:41:27,920
The objective of the analysis phase is to understand why users behave in a certain way or

362
00:41:27,920 --> 00:41:33,840
think the way they do. This is where you identify the obstacles that prevent users from carrying out

363
00:41:33,840 --> 00:41:39,840
activities and what you can do to remove the obstacles, improve efficiency, conversion and

364
00:41:39,840 --> 00:41:46,800
usability. The first thing you need to consider in your analytics is whether you have enough data

365
00:41:46,800 --> 00:41:53,920
to draw conclusions. Even if you find that you have too little data to draw large and far-reaching

366
00:41:53,920 --> 00:42:02,640
conclusions your data can suggest how you can transition the test to a larger scale and please

367
00:42:02,640 --> 00:42:07,520
don't get stuck in this phase. It's easy to find that you have spent more time analyzing than

368
00:42:07,520 --> 00:42:13,120
actually improving the internet. And you should also be aware of any seasonal variations that

369
00:42:13,120 --> 00:42:19,360
can affect the results of your review. There are peak times during the year for specific content

370
00:42:19,360 --> 00:42:26,400
and support for used on internet. For example it's common for people to forget their passwords over

371
00:42:26,400 --> 00:42:32,960
the summer holidays so demand for replacement passwords peaks when people return to work.

372
00:42:33,600 --> 00:42:39,280
This is an example of how to use your internet analytics to support the priorities of the

373
00:42:39,280 --> 00:42:45,360
editorial work on the internet. If people have extraordinary difficulties remembering their

374
00:42:45,360 --> 00:42:51,920
passwords at the end of holiday season then it's probably efficient to introduce some well-placed

375
00:42:51,920 --> 00:43:01,680
support information on how to act when one has forgotten one's password. And you should

376
00:43:02,560 --> 00:43:08,480
I think document the analytic efforts and findings based on the data you collected, processed and

377
00:43:08,480 --> 00:43:16,000
analyzed. You document the findings and lessons learned and you may choose to introduce your

378
00:43:16,000 --> 00:43:22,000
conclusions by stating the level of confidence you have in the analytics. This is wise especially

379
00:43:22,000 --> 00:43:29,120
if major decisions are likely to be taken on the basis of your analysis. Nothing wrong with

380
00:43:29,120 --> 00:43:36,320
being humble. I think this is not an exact science after all. So keep a log for future reference

381
00:43:36,320 --> 00:43:47,360
writing down which methods were used for each test etc. And also note what needs improvement

382
00:43:47,360 --> 00:43:53,840
and require another iteration through the entire analysis process. It's not unusual that you cannot

383
00:43:53,840 --> 00:44:00,240
find conclusive proof to make the case for one option or the other. And the last step in this

384
00:44:00,240 --> 00:44:07,840
process is to make improvements based on the evaluated hypothesis and the conclusions drawn

385
00:44:07,840 --> 00:44:13,680
from the analysis phase. You make a list of prioritized improvements. The improvements you

386
00:44:13,680 --> 00:44:20,080
think will matter most. It is not always easy to make those judgments. Some efforts may have

387
00:44:20,080 --> 00:44:25,840
minor impact while being an easy task and other tasks can be complex because they have external

388
00:44:25,840 --> 00:44:32,800
dependencies. No matter what, now at least you have a list of things to do and stuff you know

389
00:44:32,800 --> 00:44:40,560
is going to improve the user experience. An example of task in the improvement phase is to

390
00:44:40,560 --> 00:44:46,800
rearrange the positions of buttons to enhance usability. Something perhaps only beneficial for

391
00:44:46,800 --> 00:44:54,640
the user segment with smaller screens. And as I said before, you need to be able to

392
00:44:54,640 --> 00:45:00,560
do that. And after making your improvements, it's time for another iteration of the process.

393
00:45:00,560 --> 00:45:06,960
Internet analytics is not a project and it's never finished. The usefulness of each iteration of the

394
00:45:06,960 --> 00:45:12,560
analysis is that you get the chance to look critically at objectives, goals, and metrics

395
00:45:13,440 --> 00:45:20,480
to see if you think they still prove useful. Most likely over time you will revise, supplement,

396
00:45:20,480 --> 00:45:29,600
and verify them. And that's it. Something's up. I hope you got some great takeaways from this talk

397
00:45:29,600 --> 00:45:38,640
and to repeat the most important points I would like to pick these for. Try to set some measurable

398
00:45:38,640 --> 00:45:44,560
goals with your internet. It makes using Matomo more self-explanatory since you have conversions

399
00:45:44,560 --> 00:45:52,160
to indicate something of value inside of Matomo. Don't do internet statistics, do analytics.

400
00:45:53,360 --> 00:45:59,200
And that entails looking for actionable insights. You should use both qualitative

401
00:45:59,200 --> 00:46:07,680
and quantitative methods. They complement each other. Don't measure activity, focus on results.

402
00:46:07,680 --> 00:46:14,160
Good luck. And the consoling thought is that most other organizations also seem to suck at

403
00:46:14,160 --> 00:46:24,480
internet analytics. Let's do better. Any questions? Then you find me in the conference chatroom.

404
00:46:24,480 --> 00:46:38,080
If you'd like, you could also email me. Thank you for your attention.

405
00:46:38,080 --> 00:46:45,120
Thank you everyone for joining this session. As Markus mentioned, you can continue the conversation

406
00:46:45,120 --> 00:46:51,520
in the chat if you have further questions. I also wanted to let you know that the next round of

407
00:46:51,520 --> 00:46:58,080
sessions will start at 1 p.m. after the lunch break. So we hope to see you then.