Episode 51

Leading Without Wireless Bias – Ulf Seijmer on LoRaWAN, Cellular & IoT Strategy

Ulf Seijmer explains that leading a successful wireless IoT company requires staying close to market needs, rapidly adjusting strategy, and avoiding attachment to any single technology.

He describes how his expectations for LPWAN market splits were wrong, noting that PropTech has become dominated by LoRaWAN rather than cellular due to sensor cost economics and the advantages of owning the network.

Seijmer contrasts his roles at AKKR8, which focuses on cellular LPWAN devices, and Induo, which must remain technology-agnostic and select from LoRaWAN, cellular, satellite, and BLE based on each use case.

He emphasizes the importance of telling customers “no” when a chosen technology will not scale, and of using proofs of concept to uncover unexpected value, illustrated by a water-tap monitoring project that revealed facility issues like broken lights rather than just usage data.

Overall, he argues that long-term success comes from solution focus, honest guidance on trade-offs, and designing systems that can evolve beyond narrow, siloed applications.

Ulf on LinkedIn

Induo

AKKR8

  • Helium Global IoT Coverage - Want to know if Helium coverage exists where you need it? Check out this map!
  • Helium Foundation - The Helium Foundation's IoT Working Group (IOTWG) has generously provided support for the first 6 months of shows, please go check them out and consider using the Helium LoRaWAN as a primary or backup on your next deployment. With over a quarter million gateways deployed worldwide, it's likely that you have and can use Helium coverage.
  • Support The Show - If you'd like to support the MetSci Show financially, here's where you can donate on a one-time or an ongoing basis. Thank you!
  • MetSci Show - If you'd like to use our IoT or AI Data Value calculators, or you'd like to contact me, the MetSci Show site is the best way to do it.
  • MeteoScientific Console - Use LoRaWAN - The MeteoScientific Console allows you to use LoRaWAN today. As long as you have Helium coverage (and you probably do, about 90% of populated areas in the world have a gateway within 2 miles), you can onboard a sensor. You can always check coverage at https://explorer.helium.com and switch to the "IoT" tab in the top right.

Transcript
Speaker:

Today's guest on MeteoScientific's.

2

:

The Business of LoRaWAN

3

:

is Ulf Seijmer, Innovation leader

at Induo and co-founder and CTO of AKKR8.

4

:

Ulf brings more than 30 years

of experience in wireless and telemetry.

5

:

And in this conversation,

we focus on what it actually takes to lead

6

:

a wireless focused IoT company in a fast

changing market.

7

:

We talk about his guerrilla tactics

approach to leadership, listening closely

8

:

to what the market really needs

and adjusting quickly

9

:

instead of forcing a favorite technology.

10

:

He explains why LoRaWAN has become

dominant in Proptech, largely

11

:

driven by sensor cost and the ability

to build and control your own network.

12

:

We also dig into network economics,

including why the real cost of cellular

13

:

or LoRaWAN

14

:

often comes down to infrastructure,

man hours and the SLA you promise.

15

:

And toward the end, we explore

how IoT data and AI naturally fit together

16

:

with a real world

17

:

example of AI detecting abnormal

air quality patterns from sensor data.

18

:

This episode is sponsored

by the Helion Foundation

19

:

and is dedicated

to spreading knowledge about LoRaWAN.

20

:

If you'd like to try to image

publicly available global LoRaWAN

21

:

for free and support this show,

sign up at sideshow Slash Console.

22

:

Now let's dig into the conversation

with all the cyber.

23

:

Oh, thanks so much for coming on the show.

24

:

Super excited to have you here.

25

:

Thank you. Very excited to be here.

26

:

So unlike a couple of my guests,

you've got a lot of leadership experience.

27

:

A lot of folks come on here are really

engineering heavy and kind of RF heavy.

28

:

You've got a couple of different hats

you wear.

29

:

I thought we'd start off with.

30

:

How do you lead a good wireless

IoT company?

31

:

Are there anything

is there anything special about it?

32

:

Well, I think the special skill

that you have to adopt is that you

33

:

you have to listen to

what the market really needs,

34

:

and you have to adjust very fast

35

:

because the market is ever changing

36

:

and the opportunities are out there.

37

:

But you shouldn't

bang your head to the wall trying to

38

:

to work with your solution

that you believe in.

39

:

If if the market is asking for it

for something else.

40

:

So I call this perhaps not the best word,

but we try to run this

41

:

with the guerrilla tactics

so that we are we find what's going on

42

:

on the market and, and we go there and,

and do our best to, to meet the needs.

43

:

And then we try to regroup and

44

:

be fast on, on the next board, so to say.

45

:

Let's see, how do you find out

what's going on in the market,

46

:

or is it just very straightforward, or

do you have a structured way of doing it?

47

:

I'm fortunate to have been in the market

for 30 plus years, and,

48

:

I do have a very, very good network

and I'm super grateful

49

:

every day

that they bring along opportunities.

50

:

And I, I also try to, to network as much

as I can to find new opportunities.

51

:

And I also write a lot.

52

:

So I try to cover

what's going on, try to catch the trends

53

:

and try to be there in the forefront of

of what's going on.

54

:

Sure, if you think over the last year,

are there any changes that were unexpected

55

:

where you're like, oh, I went into 2025

thinking, we're going to do this.

56

:

And then we ended up doing doing that.

57

:

I would say that if we

if we back the tape,

58

:

if we wanted to look at it

from a, from a higher perspective.

59

:

So one one thing that, that I envision

like five years ago

60

:

was what markets would be dominated by

61

:

a certain

technology, I mean cellular level

62

:

and what market would be dominated

by LoRaWAN.

63

:

And, well, it panned out different that I

64

:

then I thought

and I said sometimes it does.

65

:

Yeah, sometimes it does.

66

:

And as you grow

it's important to still keep learning.

67

:

Right.

68

:

So I if I would have told my five year

69

:

younger version of me,

I would have thought differently

70

:

earlier to to what market

to address with certain solution.

71

:

Oh in what way was there

like a specific market that you're like,

72

:

oh, I missed that one,

or I should have hit that one harder.

73

:

I think that I saw that the telcos

were going to address the LP violence

74

:

in facilities in these kind of,

75

:

I say mass deployments, LP van indoors.

76

:

But the Proptech

77

:

market, this is almost just LoRaWAN today.

78

:

So that's one thing I,

I sort of see coming,

79

:

but I didn't see that

it would be such a definitive

80

:

market

share that there's a lot of LoRaWAN and,

81

:

and very, very little of the, of the rest

in, in the Proptech market.

82

:

As an example,

Proptech is property technology.

83

:

I just wanna make sure I'm. Yes.

84

:

Yeah. Exactly. Yes.

85

:

And why do you think LoRaWAN did so well

that they just do a good job marketing?

86

:

Did it just work better?

87

:

I know you've been pretty clear

about being agnostic towards technology.

88

:

What do you think happened what happened?

89

:

I, I think that the cost per sensor

90

:

is what what drives

the deployment basically.

91

:

And the ease of I mean, the thing is that

if you build your own network,

92

:

you can always address the

the white spots or the black spots,

93

:

and you can build the network

as you go along and as soon as you wish.

94

:

Right.

95

:

So that's easier to address.

96

:

If you own your own network,

then you you know

97

:

exactly what you're going to get and,

and what the cost is going to be.

98

:

And the big driver in this property, tech

deployments

99

:

is the cost of, of the sensors.

100

:

So the network is just,

a small part of the portion.

101

:

There are

102

:

few alternatives in the cellular market

that can be deployed at the same cost.

103

:

So that's I would say a main driver.

104

:

Oh, interesting.

105

:

So I mean just comes down to, to business,

which makes sense.

106

:

I mean, we're in business to make money.

107

:

If you can make more money by spending

less money, it they usually works out.

108

:

Yeah.

109

:

I think I've seen the argument

that cellular costs versus LoRaWAN

110

:

costs are about the same.

111

:

And despite LoRaWAN appearing to be,

quote, free, there are certainly costs

112

:

associated with it.

113

:

But it sounds like

the sensors themselves are.

114

:

What are the the difference, at least

as far as you can see it in terms of cost.

115

:

Yeah.

116

:

In these kind of project.

117

:

Yes, I, I've also done these kind of

118

:

calculations.

119

:

What what costs more the, the public

network or your, your private network.

120

:

And, and basically I would say the,

the main cost of cellular networks,

121

:

you would think that it would be

the cost of the frequency, but it's rather

122

:

the cost of running a wireless network

is, is the rest.

123

:

It's the man hours. It's

the infrastructure.

124

:

It depends on what kind of

125

:

SLA you want to build your network around.

126

:

If you want one minute

127

:

reporting time and,

everything has to be online all the time.

128

:

Or do you accept occasional values?

129

:

A couple of times per day?

130

:

So it really depends.

131

:

But I would say the large cost is the

132

:

is it man hours when you want high SLA.

133

:

Okay.

134

:

So getting that high SLA you just build in

like hey there's going to be extra cost

135

:

because if something goes down

we have to hit it right away.

136

:

Yeah.

137

:

Is that is it

138

:

the general thinking or is there

something else there in terms of cost?

139

:

Well, now I would say that's a main driver

for a wireless network.

140

:

Is the, the infrastructure in in LoRaWAN

as is only

141

:

if you want to build your own network.

142

:

It's it's a bit cheaper

than, building a cellular network.

143

:

Of course, the cellular networks

has the upside of their own frequency.

144

:

And the downside

is that the frequency cost.

145

:

But if you build a large network,

you can just forget about the

146

:

the cost of the spectrum,

because that's not that's not

147

:

going to make any difference in the end

anyway.

148

:

Got it. And how big is a large network?

149

:

Is that kind of over square miles

or number of sensors?

150

:

What does it mean?

151

:

I would say it's like a region wide Lora

one or wireless network.

152

:

So if you build for it, for a region,

153

:

if you build for a campus

or a built for a residential area,

154

:

it's not that.

155

:

It's if you build an operator

grade network for for a region.

156

:

Right, okay.

157

:

Now you're at two different companies.

158

:

It into an accurate.

159

:

What are the decisions

160

:

that you make in each at the other

other side never has to face?

161

:

I can start with the easy answer.

162

:

Inaccurate.

163

:

We develop devices

for the cellular LP bands

164

:

so we don't have to consider anything

related to to Lora.

165

:

In India,

we're building a range of products

166

:

that will meet the market needs.

167

:

So we we have to address

168

:

all the technologies that are interesting

from our customer's perspective.

169

:

It could be satellites,

it could be Lora, it could be

170

:

Bluetooth low energy and and cellular.

171

:

We're trying to find the best device

to meet a certain

172

:

need rather than to preach

a certain technology.

173

:

Okay.

174

:

And then being solution focused like that

is it sounds like that's what has kept

175

:

you kind of wireless agnostic where like,

hey, look, I can't get tied to one thing.

176

:

I just have to solve the problem.

And the best way to solve it.

177

:

Yeah. You've called yourself two headed.

178

:

I think both in our conversations

in and in, like, what do you mean by that?

179

:

Yeah.

180

:

Well, sometimes you have to argue

for the cellular networks.

181

:

Sometimes you have to argue for the

for the other networks.

182

:

I mean, in the end, I think that there is

perhaps a big focus on or on

183

:

what is the wireless bearer of the,

of the information.

184

:

Right?

185

:

So I, I'm just trying to, to find out

186

:

basically

what is best for a certain deployment.

187

:

And the funny thing is that

most of the times

188

:

when people contact me,

they have already decided.

189

:

So they already done the the homework,

or they're in a region

190

:

where they have a superior

local LoRaWAN city network.

191

:

And, and I want to go with that

because they want to support their

192

:

their local operator. Sure.

193

:

Or they want something else.

194

:

So do you think they're calling you

just like to give themselves permission

195

:

or they're calling you like, hey,

I already know what I want.

196

:

Just roll this out or something else.

197

:

Good question.

198

:

Yeah, well,

I'm often called as as the wireless expert

199

:

and I'll,

I'll try to support them in the idea.

200

:

I think that as long as you can,

if a customer has an idea

201

:

and it sounds reasonable,

it's better to to support them

202

:

and try to to find something

that can can meet their needs.

203

:

But also when you really have to say

no, when it will not work.

204

:

I, I'm not confident that

205

:

the problems

that you solve early on in a customer

206

:

journey is a problem

that will not scale later on.

207

:

If you can't say no to to small problem,

the the small problem

208

:

will become a big problem.

209

:

So you have to let them know

210

:

what are the benefits

of certain technology.

211

:

What is the potential problem

that will occur later on?

212

:

And that's super important,

not just to see

213

:

everything as an opportunity,

but also to see that the opportunity

214

:

is to solve it right,

not just to to solve it by any means.

215

:

Oh, interesting.

216

:

And how often do you think you have to say

no to a customer?

217

:

How often do you think

you have to guide them to

218

:

to do something

that maybe they didn't expect to do?

219

:

Is that kind of every single interaction

220

:

or to some customers come in, customers

come in and nail it and they get it right.

221

:

I think that the there are like

222

:

perhaps one

fifth of all customer discussions

223

:

where they can benefit from support,

and it could be just support that,

224

:

you know, if you're building a network,

you have to consider this aspect

225

:

because it's going to be potentially

a problem later on.

226

:

But at the same time,

perhaps you don't have to solve it.

227

:

By the I mean in the beginning of your

customer journey, because the problem is

228

:

that a customer would like to solve

something, would like to test something,

229

:

and it's not always reasonable

to solve it.

230

:

In the first

231

:

interaction with the customer, we can

we can talk about potential problems.

232

:

But I also prefer that they rather test

something and and evaluate something

233

:

and then later

on adjust to my recommendation.

234

:

So it's not that it has to be firm and

but you have to understand the limits now.

235

:

Interesting.

236

:

So they'll come to you the problem

237

:

and you'll say, hey,

here's what I think you should test

238

:

so that you can kind of

figure it out yourself and then come back

239

:

and we can walk through the rest of it

together.

240

:

I have a good story when it comes to

to proof of, of concepts.

241

:

This isn't my case, but I refer to it

in in discussions from time to time.

242

:

A company that develops water taps

243

:

for public environments

like airports, hospitals and so on.

244

:

They they wanted to measure

how much the water taps were being used.

245

:

And they had in an airport

as the first installation

246

:

and I believe they had Laura,

Laura sensors installed

247

:

and they tested everything

248

:

when they did the first proof of concept,

and it seemed to work just fine.

249

:

So they went back. This was in Denmark.

250

:

So they went back to Sweden,

251

:

woke up the next morning,

and they logged into their IoT platform.

252

:

And they saw that, you know,

they all water taps were being used,

253

:

but there was one that hadn't

been reporting throughout the night.

254

:

So they thought that, okay,

perhaps this one isn't really working.

255

:

Perhaps

this sensor isn't isn't working at all.

256

:

So they asked a local janitor to go out to

to this restroom and and open the door.

257

:

And he found out that the reason

the water wasn't being used

258

:

was that the the lights

were not functioning at the toilet.

259

:

So no one wanted to go into that

toilet and just, you know, close the door

260

:

and do whatever they had to do in the in

total darkness.

261

:

So no one use the water tap.

262

:

The thing is that this customer,

their customer, they,

263

:

they wanted to build a system

where they could monitor water consumption

264

:

in water tops,

but they eventually found out that

265

:

this is also a system

that can indicate other things.

266

:

So if you feed back to the cleaning staff,

to the janitors

267

:

and the local caretakers, they can monitor

268

:

these abnormalities and go

check on on the situation in the room.

269

:

Perhaps it's not clean,

perhaps it's the light is out and so on.

270

:

So so that's what I think is something

271

:

that captures

the idea of a proof of concept that

272

:

don't limit yourself in

in what your systems can achieve.

273

:

Have a good idea of what you want to do,

274

:

test it, understand

275

:

what may need to be addressed later on.

276

:

But if you box too closely

277

:

on on the problem,

you won't see the whole picture.

278

:

Interesting.

279

:

And then I'm just curious, as a

as a business owner is sold in a couple

280

:

places, was that something

where they were immediately

281

:

sold into like, hey,

let's get some light sensors on there,

282

:

or like kind of the next thing, or did

the customer, you know, require that?

283

:

And there was kind of very minimal

selling to do.

284

:

I'm not sure actually,

because in this case

285

:

it was just about the water tops.

286

:

So, I'm not sure if they,

if they ever included

287

:

other systems into to the same solution.

288

:

Possibly they did possibly date.

289

:

It was like a nice lab, like a.

290

:

Yeah.

291

:

Yeah.

292

:

But that

that could be also on the downside,

293

:

they could be that they have designed

an IoT system that only addressed

294

:

a vertical and wouldn't interact

with other kinds of systems.

295

:

And, and that's quite common as well.

296

:

You build something for a solution

that won't talk to another solution.

297

:

And and then you built

your, your silos and

298

:

you leave happy in your silo,

but you don't look sideways.

299

:

And do you think

300

:

I will start to change that

just because it's such a powerful piece

301

:

to say, hey,

302

:

here's all my data, go take a look at this

and tell me what else is out there,

303

:

or is that do you think the IoT

kind of general landscape is just like,

304

:

yeah, I've heard enough about this

AI stuff.

305

:

I'm just going to do my own thing.

306

:

I think that we need AI,

and we need AI to find these patterns

307

:

on and understand

what's going on, and and find them fast.

308

:

I mean, I did a proof of concept

a year and a half ago

309

:

where we were measuring the air quality

in a conference room and, and at a certain

310

:

time in the morning, we had a spike

of different compounds in the air.

311

:

And I spotted this.

312

:

It was just a small spike,

but we had a monitoring, and,

313

:

we we got an analysis and the reading,

314

:

the sensor data, it immediately

detected that it couldn't be humans.

315

:

It had to be something that the humans

put out in the air, basically.

316

:

So it couldn't just be, you know,

a morning meeting that that raised

317

:

or all the different values

to, to this limit.

318

:

And we discovered

that it was actually the cleaning company

319

:

that were cleaning the conference room

at a certain time in the morning,

320

:

but they were using the Covid,

cleaning detergent.

321

:

And they were just very, very aggressive.

322

:

Yeah. Sounds like. Yeah, yeah, yeah.

323

:

But the good thing was that

I spotted this quite, quite fast.

324

:

We were able to tell the, the,

325

:

the cleaning company

that they were probably using very,

326

:

very strong things to, to clean the room,

the dirt and solvents.

327

:

Yeah. Exactly.

328

:

So they, they spoke to, to the lady

who was cleaning and she addressed that.

329

:

And we got a better working environment

for, for, for her.

330

:

So I mean, it's, it's

about finding those patterns

331

:

the and and understand why they're there

and what they represent.

332

:

I mean IoT

333

:

generates loads

334

:

of data and, and that's what I want

for, for breakfast.

335

:

So I mean, it's a it's

a natural collaboration between those two.

336

:

It does seem like a natural fit.

337

:

And I think all of us at some point

have had this insight like,

338

:

oh, that IoT is that is

based on the AI sensory input system.

339

:

I think some of the people

had that insight long before I did,

340

:

but I know, I know when it hit me,

I was like,

341

:

oh my God, this thing is

this thing is amazing.

342

:

Yeah, yeah, I think it's amazing.

343

:

And I hear people looking at the eyes,

344

:

you know, it gets all the attention

while it really doesn't.

345

:

I think it's about building

powerful ecosystems.

346

:

And AI is a tool for making order

out of a data mess.

347

:

And the IoT is good at producing data.

348

:

So they they need each other

and they will be scaling together.

349

:

Read.

350

:

Well, thanks so much for carving out time

from your afternoon.

351

:

I appreciate you coming on

and talking to us about leadership and IoT

352

:

and and a little bit AI at the end.

Thank you. Thank you very much.

353

:

That's it for

this episode of The Business of LoRaWAN.

354

:

If you want to go deeper

and actually deploy devices,

355

:

the MeteoScientific

Console is the fastest way to do that.

356

:

And honestly, it's

also the best way to support the show.

357

:

When you use the console, you're not just

listening, you're participating

358

:

in the same real world LoRaWAN work

we talk about here every week.

359

:

You can get started with the free trial

at MeteoScientific.com.

360

:

Huge thanks to the sponsor of the show,

the Helium Foundation,

361

:

for supporting open LoRaWAN

infrastructure worldwide.

362

:

Check them out at helium.foundation

and if the show has been

363

:

useful, a quick rating or review

on Apple Podcasts or wherever you listen.

364

:

This really helps

365

:

people find it and helps the show grow

so we can help more people.

366

:

I'm Nik Hawks with MeteoScientific.

367

:

I'll catch you on the next episode.

All Episodes Previous Episode

About the Podcast

Show artwork for The Business of LoRaWAN
The Business of LoRaWAN
Learn From the Pros

Listen for free

About your host

Profile picture for Nik Hawks

Nik Hawks

Incurably curious, to stormy nights and the wine-dark sea!