1
00:00:00,000 --> 00:00:03,920
Welcome back to The Deep Dive. We're here to take a whole stack of sources, docs,

2
00:00:03,920 --> 00:00:08,240
comparisons, technical papers, and, you know, just give you the shortcut to what

3
00:00:08,240 --> 00:00:09,200
actually matters.

4
00:00:09,200 --> 00:00:13,200
Right. And today we're dissecting a really interesting solution in web analytics.

5
00:00:13,200 --> 00:00:16,160
We are. For years, the default was always,

6
00:00:16,160 --> 00:00:19,470
if you wanted to know what users were doing on your site, you just went to the

7
00:00:19,470 --> 00:00:20,320
giants.

8
00:00:20,320 --> 00:00:25,600
Oh, absolutely. But those systems have become, frankly, a mess.

9
00:00:25,600 --> 00:00:30,080
It's information overload, confusing pricing, and the biggest one, a massive

10
00:00:30,080 --> 00:00:31,440
compliance headache.

11
00:00:31,440 --> 00:00:35,580
Exactly. The pressure now is all about getting good data without compromising

12
00:00:35,580 --> 00:00:36,240
privacy

13
00:00:36,240 --> 00:00:38,960
and without drowning your site in all those consent pop-ups.

14
00:00:38,960 --> 00:00:42,960
So our deep dive today is into a tool called Vince.

15
00:00:42,960 --> 00:00:47,040
It's a self-hosted platform that basically claims to solve this exact problem.

16
00:00:47,040 --> 00:00:49,440
And we've been through the GitHub docs, the landing pages,

17
00:00:49,440 --> 00:00:53,040
looking at the technical promises, how easy it is to operate, all of that.

18
00:00:53,040 --> 00:00:57,280
Right. And our mission here is to unpack how this single binary zero dependency

19
00:00:57,280 --> 00:00:57,920
tool

20
00:00:57,920 --> 00:01:03,040
makes compliance simple and gives you real insights, especially for you.

21
00:01:03,040 --> 00:01:07,040
If you're maybe a beginner to self-hosting or just tired of managing huge

22
00:01:07,040 --> 00:01:07,760
infrastructure,

23
00:01:07,760 --> 00:01:09,200
just to count clicks.

24
00:01:09,200 --> 00:01:14,880
And what we're seeing is a solution where simplicity, it's not just a feature,

25
00:01:14,880 --> 00:01:17,040
it's the core architectural principle.

26
00:01:17,040 --> 00:01:20,080
Okay. Before we get into the genius behind its lightweight design,

27
00:01:20,080 --> 00:01:23,120
we really need to thank the supporter of this deep dive.

28
00:01:23,120 --> 00:01:23,760
Of course.

29
00:01:23,760 --> 00:01:25,920
This is all made possible by Safe Server.

30
00:01:25,920 --> 00:01:29,600
Safe Server supports the hosting of innovative software, just like Vince,

31
00:01:29,600 --> 00:01:32,240
and they can help you with your digital transformation.

32
00:01:32,240 --> 00:01:36,320
You can learn more at www.safeserver.de.

33
00:01:36,320 --> 00:01:37,840
So let's set the goal for today.

34
00:01:37,840 --> 00:01:38,720
Let's do it.

35
00:01:38,720 --> 00:01:43,620
We want you to walk away understanding not just what Vince is, but why its design

36
00:01:43,620 --> 00:01:44,080
choices,

37
00:01:44,080 --> 00:01:46,880
you know, the zero dependency, the custom storage,

38
00:01:46,880 --> 00:01:50,880
why that represents a real strategic shift in how you can handle your own data.

39
00:01:50,880 --> 00:01:52,960
So let's start with a problem Vince is trying to fix.

40
00:01:52,960 --> 00:01:55,840
It's not just that Google Analytics is complex.

41
00:01:55,840 --> 00:01:58,480
The sources point out that it often gives you too much information.

42
00:01:58,480 --> 00:01:59,760
It's like data smog.

43
00:01:59,760 --> 00:02:00,480
Data smog.

44
00:02:00,480 --> 00:02:01,040
I like that.

45
00:02:01,040 --> 00:02:02,320
It paralyzes you.

46
00:02:02,320 --> 00:02:06,080
But for most people listening, I bet the biggest pain isn't even the complexity.

47
00:02:06,080 --> 00:02:07,280
It's the cost.

48
00:02:07,280 --> 00:02:09,360
Not money, but the regulatory cost.

49
00:02:09,360 --> 00:02:10,000
That's it.

50
00:02:10,000 --> 00:02:11,360
That's the real barrier.

51
00:02:11,360 --> 00:02:14,400
Compliance has gotten so complicated.

52
00:02:14,400 --> 00:02:17,840
If you're running a simple blog or a small shop, you don't want to have to hire a

53
00:02:17,840 --> 00:02:18,160
legal

54
00:02:18,160 --> 00:02:19,760
team just to manage your cookie banners.

55
00:02:19,760 --> 00:02:21,200
No, absolutely not.

56
00:02:21,200 --> 00:02:24,320
And this is where Vince makes its first big strategic move.

57
00:02:24,320 --> 00:02:27,280
So for the beginner, let's just define self-hosting.

58
00:02:27,280 --> 00:02:29,120
You control the server.

59
00:02:29,120 --> 00:02:30,480
You control the data.

60
00:02:30,480 --> 00:02:31,200
That's fundamental.

61
00:02:31,200 --> 00:02:31,520
Okay.

62
00:02:31,520 --> 00:02:36,480
But Vince has built from the ground up to be compliant with GDPR, CCPA, PCR,

63
00:02:36,480 --> 00:02:39,120
all the big ones, right out of the box.

64
00:02:39,120 --> 00:02:43,520
Wait, how can a tool just claim compliance without making the user spend months

65
00:02:43,520 --> 00:02:45,840
configuring legal notices and settings?

66
00:02:45,840 --> 00:02:47,600
That sounds a little too good to be true.

67
00:02:47,600 --> 00:02:50,000
It's a really elegant architectural decision.

68
00:02:50,000 --> 00:02:51,760
The documentation is very clear.

69
00:02:51,760 --> 00:02:53,840
Vince uses no cookies.

70
00:02:53,840 --> 00:02:54,960
No cookies at all.

71
00:02:54,960 --> 00:02:55,600
None.

72
00:02:55,600 --> 00:02:58,000
And it ensures 100% data ownership.

73
00:02:58,000 --> 00:03:01,920
By killing tracking cookies entirely and storing all the data on your server,

74
00:03:01,920 --> 00:03:04,960
it just sidesteps the need for most of those consent pop-ups.

75
00:03:04,960 --> 00:03:05,600
Ah, okay.

76
00:03:05,600 --> 00:03:09,760
So if you're not tracking individuals across different websites or sessions

77
00:03:09,760 --> 00:03:12,400
with those unique identifiers,

78
00:03:12,400 --> 00:03:17,040
the whole compliance burden just shrinks dramatically.

79
00:03:17,040 --> 00:03:18,000
It falls off a cliff.

80
00:03:18,000 --> 00:03:22,480
So instead of trying to manage consent for all this complex tracking,

81
00:03:22,480 --> 00:03:26,480
they just decided not to track in a way that requires that kind of consent.

82
00:03:26,480 --> 00:03:29,280
That is a massive operational shortcut.

83
00:03:29,280 --> 00:03:32,240
It means you can literally get rid of those annoying cookie pop-ups

84
00:03:32,240 --> 00:03:34,000
that just ruin the user experience.

85
00:03:34,000 --> 00:03:34,400
Exactly.

86
00:03:34,400 --> 00:03:35,760
It removes friction for everyone.

87
00:03:35,760 --> 00:03:36,240
Yeah.

88
00:03:36,240 --> 00:03:39,040
But there's another side to the operational equation,

89
00:03:39,040 --> 00:03:40,640
and that's just the ease of management.

90
00:03:40,640 --> 00:03:40,960
Right.

91
00:03:40,960 --> 00:03:44,000
For a beginner who's just dipping their toe into self-hosting,

92
00:03:44,000 --> 00:03:50,000
the idea of managing databases and external software, it's terrifying.

93
00:03:50,000 --> 00:03:52,720
And that's where Vince comes in with what they call painless deployment.

94
00:03:52,720 --> 00:03:56,240
The key technical detail here is that it ships as a single binary

95
00:03:56,240 --> 00:03:57,920
with no external dependencies.

96
00:03:57,920 --> 00:03:58,480
Yes.

97
00:03:58,480 --> 00:04:02,080
Let's make sure we really break down what single binary actually means for you,

98
00:04:02,080 --> 00:04:04,000
the person who has to manage this thing.

99
00:04:04,000 --> 00:04:06,080
It means simplicity, defined.

100
00:04:06,080 --> 00:04:07,920
A normal web app needs what?

101
00:04:07,920 --> 00:04:13,120
An operating system, a web server, a programming language like Elixir or Node,

102
00:04:13,120 --> 00:04:15,040
and a database like PostgreSQL.

103
00:04:15,040 --> 00:04:16,640
A whole stack of things to manage.

104
00:04:16,640 --> 00:04:17,680
A whole stack.

105
00:04:17,680 --> 00:04:23,200
A single binary means the entire application is just one file, one single executable.

106
00:04:23,200 --> 00:04:26,720
You run that one file and the whole analytics engine is live.

107
00:04:26,720 --> 00:04:30,000
So the setup and the maintenance effort just plummets.

108
00:04:30,000 --> 00:04:32,560
There's only one thing you ever need to patch or update.

109
00:04:32,560 --> 00:04:34,960
Fewer moving parts means fewer things can break.

110
00:04:34,960 --> 00:04:38,320
Okay, so if we're looking at this as a strategic choice,

111
00:04:38,320 --> 00:04:42,640
it sounds like Vince is sacrificing features or power for that simplicity.

112
00:04:42,640 --> 00:04:44,080
Is that the trade-off here?

113
00:04:44,080 --> 00:04:45,840
That is the critical question.

114
00:04:45,840 --> 00:04:48,560
If you get to zero dependency, what did you have to give up?

115
00:04:48,560 --> 00:04:51,360
And to answer that, we have to look at the competition

116
00:04:51,360 --> 00:04:54,080
because Vince was heavily inspired by other projects.

117
00:04:54,080 --> 00:04:55,360
So give us that contrast.

118
00:04:55,360 --> 00:04:58,320
I mean, anyone who's fought with setting up a database cluster

119
00:04:58,320 --> 00:05:01,920
just for a simple web counter knows how good one file sounds.

120
00:05:01,920 --> 00:05:05,360
OK, let's take Plausible Analytics, which Vince originally used as a model.

121
00:05:05,360 --> 00:05:10,080
Plausible is a fantastic privacy-focused tool, no question.

122
00:05:10,080 --> 00:05:14,560
However, to self-host it, you typically need to manage Elixir for the application,

123
00:05:14,560 --> 00:05:16,880
ClickHouse for the high-speed data storage,

124
00:05:16,880 --> 00:05:19,760
and PostgreSQL for the relational database parts.

125
00:05:19,760 --> 00:05:23,600
OK, that's three major separate pieces of software,

126
00:05:23,600 --> 00:05:26,960
each with its own configuration, its own updates, its own security patches.

127
00:05:26,960 --> 00:05:31,360
It's basically a full-time sysadmin job for a small analytics setup.

128
00:05:31,360 --> 00:05:32,400
Yeah, it can be.

129
00:05:32,400 --> 00:05:34,880
Vince, on the other hand, just says forget Elixir,

130
00:05:34,880 --> 00:05:36,960
forget ClickHouse, forget PostgreSQL.

131
00:05:36,960 --> 00:05:39,760
It handles everything, all the logic,

132
00:05:39,760 --> 00:05:43,280
all the data storage internally inside that one binary file.

133
00:05:43,280 --> 00:05:47,040
So for you, the listener, installation isn't some three-stage engineering project.

134
00:05:47,040 --> 00:05:50,960
It's literally just running a script or pulling a Docker container.

135
00:05:50,960 --> 00:05:54,000
Exactly. The sources confirm how easy it is.

136
00:05:54,000 --> 00:05:56,640
For Mac or Linux, it's a simple curl script.

137
00:05:56,640 --> 00:06:00,000
You're up and running. If you like containers, it's a Docker pull.

138
00:06:00,000 --> 00:06:02,880
It just simplifies that jump into self-hosting so much.

139
00:06:02,880 --> 00:06:05,120
And it seems pretty mature for being so simple.

140
00:06:05,120 --> 00:06:08,480
It has native support for automatic TLS through Let's Encrypt.

141
00:06:08,480 --> 00:06:12,640
It does, which means securing your data with HTTPS is built in and automated.

142
00:06:12,640 --> 00:06:15,360
That's another complex system you don't have to manage yourself.

143
00:06:15,360 --> 00:06:19,360
And since you're running it yourself, you're not hitting any commercial paywalls.

144
00:06:19,360 --> 00:06:24,640
The docs really emphasize this. Unlimited sites and unlimited events.

145
00:06:24,640 --> 00:06:28,790
Your only limit is the resources of your own server, not some vendor's subscription

146
00:06:28,790 --> 00:06:29,200
plan.

147
00:06:29,200 --> 00:06:33,360
So back to my question about sacrificing capability. It's lean, but what can it

148
00:06:33,360 --> 00:06:34,080
actually track?

149
00:06:34,080 --> 00:06:38,480
It tracks everything a modern website owner really needs. It's not just page views.

150
00:06:38,480 --> 00:06:42,080
It does goal conversion tracking so you know when users sign up or buy something.

151
00:06:42,080 --> 00:06:42,800
That is important.

152
00:06:42,800 --> 00:06:48,720
It handles custom events, outbound links, file downloads, and even 404 error pages,

153
00:06:48,720 --> 00:06:52,080
which is critical for site health and SEO.

154
00:06:52,080 --> 00:06:56,640
So the core functionality is all there, but the operational headache is just gone.

155
00:06:56,640 --> 00:06:57,120
Pretty much.

156
00:06:57,760 --> 00:06:59,680
This brings us to the architecture.

157
00:06:59,680 --> 00:07:02,640
We mentioned Vince started as a Go port of Plausible,

158
00:07:02,640 --> 00:07:05,360
aiming specifically for that self-hosting crowd.

159
00:07:05,360 --> 00:07:06,480
That lineage is key.

160
00:07:06,480 --> 00:07:11,600
They looked at a successful analytics tool and just rebuilt the engine in Go,

161
00:07:11,600 --> 00:07:14,720
which is a language that's perfect for creating single binaries.

162
00:07:14,720 --> 00:07:17,200
And the focus on self-hosting is so complete

163
00:07:17,200 --> 00:07:21,340
that the tracking script is actually designed as a drop-in replacement for Plausible

164
00:07:21,340 --> 00:07:21,840
script.

165
00:07:21,840 --> 00:07:24,880
Yes, which makes migration incredibly easy

166
00:07:24,880 --> 00:07:28,240
if you're already using Plausible's self-hosted version.

167
00:07:28,240 --> 00:07:31,040
Okay, now let's tackle that secret sauce you mentioned,

168
00:07:31,040 --> 00:07:34,720
the thing that allows it to be fast without a massive database.

169
00:07:34,720 --> 00:07:39,680
The sources mention compressed roaring bitmaps using columnar storage.

170
00:07:39,680 --> 00:07:41,840
Let's simplify that.

171
00:07:41,840 --> 00:07:46,000
Okay, think of a traditional database like a giant spreadsheet or a ledger.

172
00:07:46,000 --> 00:07:50,000
To count something, you have to read through every row line by line.

173
00:07:50,000 --> 00:07:52,480
It gets slow really fast with millions of events.

174
00:07:52,480 --> 00:07:53,200
Right.

175
00:07:53,200 --> 00:07:56,280
Columnar storage is like taking that spreadsheet and cutting it into vertical

176
00:07:56,280 --> 00:07:56,880
scripts.

177
00:07:56,880 --> 00:07:58,560
All the page views are stored together,

178
00:07:58,560 --> 00:08:00,720
all the timestamps are stored together, and so on.

179
00:08:00,720 --> 00:08:02,800
That already sounds way more efficient for a running report.

180
00:08:02,800 --> 00:08:03,440
It is.

181
00:08:03,440 --> 00:08:05,440
Now, you add the roaring bitmaps.

182
00:08:05,440 --> 00:08:08,720
Imagine you need to know which users visited a specific page.

183
00:08:08,720 --> 00:08:11,920
Instead of storing a huge list of every single user ID,

184
00:08:11,920 --> 00:08:15,440
a roaring bitmap is like a super compressed index card

185
00:08:15,440 --> 00:08:17,600
that just flags where the relevant data is.

186
00:08:17,600 --> 00:08:20,720
So it's not storing all the raw data in a traditional way,

187
00:08:20,720 --> 00:08:24,800
it's storing a highly efficient compressed map of the data.

188
00:08:24,800 --> 00:08:25,600
Exactly.

189
00:08:25,600 --> 00:08:29,120
A map that lets it pull reports and metrics almost instantly

190
00:08:29,120 --> 00:08:32,560
without needing a huge complex query engine like ClickHouse.

191
00:08:32,560 --> 00:08:36,240
That addresses the speed and the efficiency.

192
00:08:36,240 --> 00:08:36,720
It does.

193
00:08:36,720 --> 00:08:40,720
It gives you extreme query speeds and very efficient disk usage

194
00:08:40,720 --> 00:08:42,880
all managed inside that single file.

195
00:08:42,880 --> 00:08:46,320
And the tracking script you put on your website, less than one kilobyte.

196
00:08:46,320 --> 00:08:48,880
Which means your site performance won't take a hit

197
00:08:48,880 --> 00:08:50,240
just because you're running analytics.

198
00:08:50,240 --> 00:08:51,760
The zero impact, basically.

199
00:08:51,760 --> 00:08:54,960
So if we look at the feature comparison table and the sources,

200
00:08:54,960 --> 00:08:57,200
the different strategies become really clear.

201
00:08:57,200 --> 00:09:01,520
Plausible is built to potentially be a big hosted enterprise solution.

202
00:09:01,520 --> 00:09:04,560
Vince is laser focused on the single person,

203
00:09:04,560 --> 00:09:06,960
the single entity who wants to self-host.

204
00:09:06,960 --> 00:09:08,320
That's it, exactly.

205
00:09:08,320 --> 00:09:11,200
For the person who insists on owning their infrastructure,

206
00:09:11,200 --> 00:09:13,040
Vince gives them key advantages.

207
00:09:13,040 --> 00:09:16,960
Unlimited sites, unlimited events, zero dependency, automatic TLS.

208
00:09:17,520 --> 00:09:21,280
Plausible's architecture is excellent, but it requires that more complex setup

209
00:09:21,280 --> 00:09:23,120
because it's built for multi-tenancy.

210
00:09:23,120 --> 00:09:25,680
And we should be clear about what Vince doesn't do.

211
00:09:25,680 --> 00:09:27,200
The sources are upfront about it.

212
00:09:27,200 --> 00:09:30,800
No big enterprise features, no hosted offering,

213
00:09:30,800 --> 00:09:34,960
no multi-tenant capabilities, no advanced funnel analysis.

214
00:09:34,960 --> 00:09:36,880
And that just reinforces its mission.

215
00:09:36,880 --> 00:09:41,600
It is streamlined, high-speed, compliant analytics for one entity.

216
00:09:41,600 --> 00:09:46,000
It's built for autonomy, not for being a massive commercial service.

217
00:09:46,000 --> 00:09:48,160
It knows exactly who its audience is.

218
00:09:48,160 --> 00:09:50,720
But autonomy doesn't have to mean isolation, right?

219
00:09:50,720 --> 00:09:51,920
Even if you host it yourself,

220
00:09:51,920 --> 00:09:54,560
you still need to share data with your team or clients.

221
00:09:54,560 --> 00:09:55,760
And they thought of that.

222
00:09:55,760 --> 00:09:59,280
You can create public dashboards and unique shared access links.

223
00:09:59,280 --> 00:10:02,480
And importantly, these links can be password protected.

224
00:10:02,480 --> 00:10:05,680
Ah, so you keep total control over the data and security,

225
00:10:05,680 --> 00:10:07,600
but you can still share specific views

226
00:10:07,600 --> 00:10:09,360
without having to give people server access.

227
00:10:09,360 --> 00:10:12,160
Precisely. No need to even create user accounts for them.

228
00:10:12,160 --> 00:10:14,960
So what does this all mean for you, the listener?

229
00:10:14,960 --> 00:10:17,360
We dove into Vince looking for a self-hosted tool

230
00:10:17,360 --> 00:10:20,320
that balances speed, compliance, and ease of use.

231
00:10:20,320 --> 00:10:22,080
And what we found is an identity

232
00:10:22,080 --> 00:10:24,640
that's just defined by architectural minimalism.

233
00:10:24,640 --> 00:10:26,000
I think the takeaway is very clear.

234
00:10:26,000 --> 00:10:28,480
Vince is a cost-effective, easily deployed

235
00:10:28,480 --> 00:10:30,080
zero-dependency analytics tool.

236
00:10:30,080 --> 00:10:33,040
Its use of this compressed columnar storage

237
00:10:33,040 --> 00:10:34,560
makes it fast and efficient,

238
00:10:34,560 --> 00:10:36,320
while its core privacy stance,

239
00:10:36,320 --> 00:10:38,480
no cookies, 100% data ownership,

240
00:10:38,480 --> 00:10:41,040
gives you immediate compliance relief.

241
00:10:41,040 --> 00:10:44,000
For anyone tired of managing a bunch of dependencies

242
00:10:44,000 --> 00:10:45,440
or complex databases,

243
00:10:45,440 --> 00:10:48,080
this is just a very attractive package.

244
00:10:48,080 --> 00:10:48,640
It is.

245
00:10:48,640 --> 00:10:50,240
Which brings us to the final

246
00:10:50,240 --> 00:10:52,160
really important question for you to think about.

247
00:10:52,160 --> 00:10:54,240
If you have 100% data ownership,

248
00:10:54,240 --> 00:10:55,760
the data is local, it's private,

249
00:10:55,760 --> 00:10:57,440
it's governed only by your choices,

250
00:10:57,440 --> 00:10:59,680
how does that fundamentally change

251
00:10:59,680 --> 00:11:00,800
your analytics strategy

252
00:11:00,800 --> 00:11:02,640
compared to just relying on a third party?

253
00:11:02,640 --> 00:11:05,440
When you can treat your analytics data

254
00:11:05,440 --> 00:11:06,960
as a pure strategic asset

255
00:11:06,960 --> 00:11:08,800
that you completely control,

256
00:11:08,800 --> 00:11:11,280
what new insights or even new business models

257
00:11:11,280 --> 00:11:12,320
become possible?

258
00:11:12,320 --> 00:11:14,320
It shifts the entire conversation.

259
00:11:14,320 --> 00:11:17,360
You move away from vendor lock-in and regulatory fear

260
00:11:17,360 --> 00:11:19,120
and you move towards leveraging knowledge

261
00:11:19,120 --> 00:11:21,280
that you actually own for your own growth.

262
00:11:21,280 --> 00:11:22,720
A perfect place to leave it.

263
00:11:22,720 --> 00:11:24,480
As you think about

264
00:11:24,480 --> 00:11:27,200
where to host this kind of strategic asset

265
00:11:27,200 --> 00:11:30,080
and start on your own digital transformation journey,

266
00:11:30,080 --> 00:11:33,120
remember this deep dive was supported by SafeServer.

267
00:11:33,120 --> 00:11:35,360
They support hosting for software like Vents

268
00:11:35,360 --> 00:11:38,080
and can help with your broader digital transformation needs.

269
00:11:38,080 --> 00:11:42,560
You can learn more about them at www.safeserver.de.

270
00:11:42,560 --> 00:11:44,240
Thank you for joining us for this deep dive

271
00:11:44,240 --> 00:11:46,000
into Privacy First Web Analytics.

272
00:11:46,000 --> 00:11:47,040
We'll catch you next time.