Welcome to a new episode where we share stories from the field. For the first time, we're thrilled to welcome guests to the show!

This week, we're joined by Elena Stojmilova, Technical Lead at Open GI, and Peter Hunter, Technical Architect at Open GI, alongside our Hosts Andrea Magnorsky and Kenny (baas) Schwegler. Elena shares her personal journey and lessons learned from implementing a decentralised decision-making process within her team at Open GI, including the shift to autonomous teams and the introduction of Architectural Decision Records (ADRs).

Elena provides a candid look at the challenges and triumphs of moving from a software engineering focus to taking on full architectural decision-making authority.

Key Discussion Points

• Decentralised Decision-Making: The necessary move to create independent, autonomous teams and empower Technical Leads to make architectural decisions.
• The Mindset Shift: Moving from a coding focus to considering the broader impact of decisions, including cost and system-wide effects.
• The Power of Support: The crucial importance of technical and soft skills guidance when transitioning into a new leadership role.
• Architectural Decision Records (ADRs): The process introduced to formalise decisions, helping guide the team and ensuring accountability.
• Navigating the Advisory Forum: The challenge of managing many strong opinions initially, and the evolution toward receiving more constructive advice.
• Facilitating Advice: The techniques used to manage opinionated discussions, including asking questions to uncover the reasoning behind feedback.
• Cultural Change: How the process promoted a culture of knowledge sharing between teams and the need for architects to adapt their role from "broadcasting" to facilitating.

Hello and welcome to a new episode of Stories in facilitating

2

software design and architecture.

3

We, are very happy that we have

our first guests in our series

4

of media where we share stories.

5

Today we have with us not only, our

good companion, Kenny, who's here.

6

Hello, Kenny.

7

But also we have Elena Uba technical lead

for Open gi, as well as Peter Hunter, who

8

is the technical architect for Open gi.

9

Today we're gonna hear a story from Elena.

10

So do you wanna take

it away from us, Elena?

11

Elena Stojmilova: Hello.

12

as Andrea mentioned, I'm technically

lead in open G and we have implemented

13

and we are still using decentralized

decision making process, in our,

14

way of working in our company.

15

So I will try to make like a timeframe

how this started and how it's going now.

16

So our beginnings were like four,

almost five years ago in the time when

17

the world was changing, everything

was changing, and we also decided to

18

make some changes within our company.

19

We did a lot of changes, in the ways of

working in the technology in we, we wanted

20

to migrate to cloud and, according that

we also wanted, to change the way how we

21

are making the architectural decision.

22

Thinking in that time, it was quite a

lot, a lot changes, a lot to consume.

23

But now, and I'm reflecting on it,

maybe it's the best way because it

24

all comes and everything is changing.

25

And you're saying you started to think,

okay, it's good to try something new,

26

to try something new and everything.

27

my team was one of the teams that we

started a new program, Cloudberry,

28

in which, my team was part of it.

29

And, we, changed the ways of working,

but on another side, we wanted to create,

30

independent autonomy teams that will

make the architectural decision, will

31

be responsible for different context

from our product and will be independent

32

in the way how we implemented that.

33

So the first biggest change for me was

that the first part when the team was able

34

to make the architectural architecture

decision, it was a huge change.

35

Well, like at the beginning I was.

36

Okay, that's great.

37

We will finally be able to decide what

technology we're going to use, how

38

we'll make the things and we were like,

okay, we can try different technologies.

39

We can try this, we can try

this or, and everything.

40

But when we started actually to do it

and think it, it was, oh, it's not that

41

easy because you are your first time.

42

We were always working as a

engineer, as a software engineer,

43

and your mindset and the way of

thinking is different in that part.

44

You are mostly focused of, coding and

best, coding standards and everything.

45

If you are more experienced, you

will think like how this can be made.

46

What will be the impact of if I made this,

changed, let's say the database structure,

47

if I'm not using relation database

and everything, but when you actually.

48

Are kind of have power

to do that decision.

49

It's, yeah, first, first time.

50

Okay.

51

We can try non relation database.

52

It'll be great.

53

We'll learn something new.

54

It's something that I've always wanted

to try to do it, but, when you actually

55

start to think of it and you still,

you have like a principle that you

56

need to follow, you start to think,

oh, there is a cost there, I must.

57

Be, thinking of the cost.

58

And after that you would think like,

okay, there is other things that are

59

going to be impacted from my decision.

60

So I will change their, the way

of their, how they're working.

61

I will change some of the contracts.

62

So you need to think of them.

63

And that's where come the

reality came and, but on another.

64

But thinking now for me, that's, the most

powerful way of learning because you are.

65

Starting to see the wall picture, you're

starting to think of, the product you

66

are building, not just like one piece

of application, one piece of code.

67

You're thinking of the all, the system.

68

And, when it comes to actually making

the decision, for me the most important

69

thing is, to have the support.

70

So at the beginning, you'll not have

the skills to make that decision.

71

Technical, non-technical decision, skills.

72

And it's really important to get

some support through that process.

73

And I must say, thinking back from

that time, Pete and Andrew were the.

74

The people who really support me

and the other technical leads and

75

everyone, they were guiding us.

76

If, as I remember now, I used to come

back to, especially to Pete, like

77

with the question, should we do this?

78

What do you think is the,

is this the best way?

79

But, he was kind of changing that

way because we were work previously

80

and in that time, like it was other

people that was just trying to.

81

Give me some advices.

82

You can try this, but consider this also.

83

But see what will happen

if you choose another way.

84

And also here in injury, they

didn't go with the, this is the

85

best option, you should do this.

86

But they were just advising

us and supporting us.

87

And this support is not only for

the technical part, it's also for

88

the soft skills to kind of help

you to grow in that role because.

89

At that time I'm technical lead and

I have a team that I needed to lead.

90

And it was also, new for me.

91

I was kind of used on that change.

92

And also I needed to kind of change,

change the team, it change how the

93

team is, thinking and how the team is

performing and how the team is working.

94

And I must say that I also

get some support and advices

95

from Peter and Andrew here.

96

And that is very important, not just to.

97

Support the people into their technical

skills and what technical approaches

98

to do what architectural decisions

to be made, but also support them in

99

the way of how they're making that

and when I'm here, I will say the

100

architectural decision records that

we introduced, they are one of the

101

biggest help for me at the beginning.

102

The first one, it wasn't maybe

the best one because it was.

103

First one, and it was a big decision.

104

We made it and we were like trying to

change everything because you, you know,

105

you want to do trying something new.

106

You want to build something new

and you want to do it in the best

107

ways, like the best architecture,

principles and everything, but.

108

And we did it like we

made that first, a DR.

109

From perspective from now, it was too

much, too many decision maybe with,

110

from, taking from now, I will try

more iterative approach, doing smaller

111

decisions and doing, doing it one by one.

112

The first architecture,

advisory forum, it was.

113

Good.

114

Because in that time when you're

first starting, architectural advisory

115

forums are really proudly, like you

have a lot of people there, a lot of

116

opinions, a lot of advices and opinions.

117

A lot of people can have a strong

opinion and they will share it.

118

That is a good, but another society

can influence your decision.

119

In a way maybe that you

don't want to do it.

120

So you need to be kind of calm and to go

through all of that advices, all of that

121

options, consider each of them, but don't

let data influence your decision in a way.

122

Like, okay, he's saying that

he's the most, I know him.

123

He's, very proficient with

that and that technology.

124

So I will just follow his

advice or his opinion.

125

In architecture decision record process.

126

I, I will say now, I'm using it as a

process, as a guidance when I'm making

127

the decisions, it really helped me,

because when I'm starting to do like this

128

is we have something new to implement

and we need to make a new decision.

129

Just try starting with writing in a DR

At the beginning it was just me for,

130

as a technical lead because somehow

I wanted to get used to that practice

131

of doing the decision and also using

the ads and writing them after that.

132

We, now, when we are doing that as

a team, so we are mostly, sometimes

133

we are writing, we are calling

and starting to think like this

134

is the decision we needed to make.

135

This is the context.

136

This is, what we needed to achieve.

137

Those are our architecture principles.

138

So let's see, what are our option?

139

We are going one by one of

the, each of the option.

140

As a team, we are deciding, doing, need

any spiking or additional investigation,

141

for any of those options to get more

knowledge or to try it and see it.

142

So we are.

143

They are, guiding comes

in creating a new spikes.

144

And after that, when everything

is finished, we are consider

145

all the results as a team also,

and we are making that decision.

146

Now in this, like after

four years also, I'm.

147

Trying to get more, advices during

the process of the decision when

148

it's needed from the architect.

149

I must say at the beginning,

it wasn't like that.

150

First we were making the decision

and after that we were just sharing

151

conduct, actually advisory forum, and.

152

Yeah.

153

After that we received that advice.

154

But now, when we are practicing this,

it's more efficient when you're including

155

other people like architects or other

people, stakeholders, that can help you

156

in the process of making the decision.

157

So, I'm trying to include some of our

architects when I'm thinking of all

158

options, like writing some, high level

diagram and asking them for any advice

159

or any thoughts or any, uh, or some of

their opinions because I might miss some

160

technical knowledge, or I miss some.

161

I'm maybe not aware of any context

knowledge, or I missed something

162

about the product and everything.

163

And that's really, helped me.

164

It's helping me.

165

And, when you are writing the architecture

decisions record and everything,

166

maybe the most common question is how

you are making sure that the people

167

will, do the things that they're.

168

Writing on, like when they're writing

the architecture decision record,

169

making sure that they will do that.

170

For me it's most important to have,

people who has like that culture

171

of people that have accountability

and they have a sense of ownership.

172

Yes.

173

When you are starting, you must

choose at the beginning people

174

who have that soft skills also.

175

Because it's not just that you need to

write idea, it's your responsibility

176

and you're accountable that, what is

writing in the idea you will make it.

177

There were some key situation in the

past also for me, when some of the

178

decisions we have made have changed.

179

Yes, we, it's, that's can, happen and that

will happen because over time you can.

180

came like, oh, maybe this

is not the right decision.

181

Maybe we should do it.

182

We should, we should do

it in the different way.

183

You can, change, you can, change it.

184

You can supersede the previous idea

and write a new idea and that's okay,

185

but you need to make the culture

that you will write a new idea in.

186

Do that.

187

How the ideas are helping us as a

team is they're sharing because,

188

on the architecture advisory forum,

all the teams are sharing the ideas.

189

We have a kind of culture

of shared knowledge.

190

Because in the beginning, when this wasn't

like, a process, each of the team was

191

working installation and we were mostly

focused on, this is our context, this

192

is our work and this is our context.

193

Yeah, you can hear from someone like,

but in informal conversation, some

194

of the team is doing this on some

presentation, but it wasn't formalized.

195

But now when you are in architecture

advisory forum, you can hear what

196

other teams are doing, what are

the decision they making, what are

197

the technologies they're using?

198

And some, sometimes you can come to them

and ask them for help and, advising.

199

Sometimes you can reuse,

some of their implementation.

200

And that is a really big, benefit

for me because sometimes you see,

201

okay, okay, they have tried this.

202

Can we ask them to share their knowledge

and we can see and consider we can try

203

this also in, in our, in our context.

204

Andrea Magnorsky: It's great to hear that

from your direct experience about how

205

the advice form you with the knowledge

sharing because, and you also mentioned

206

that they, the, especially at first,

those seem quite crowded and, also

207

that it was kind of hard to, right.

208

I need to keep my head on and so, so

that, that's, really nice to have a, a,

209

your, your first person account of that.

210

Do you wanna, like, tell us a little

bit more about that, bit, about the

211

bit of, the, that first time and maybe

how it's, how you grew from, you know,

212

maybe getting more confident with it.

213

Elena Stojmilova: Yeah, at the beginning,

especially on the first a D area,

214

there was a lot of comments during

the architectural advisory forum, the

215

meeting itself, but I must say here was.

216

Pete and Andrew, and they, they were

good facilitators and it's really

217

important to have a good facilitators

during that meetings, because.

218

There will be situation when no one

will have opinion and there will be

219

situation when everyone wanted to have

their opinion and they will share it.

220

We have like a, on the first aid, lots

of, a lot of comments and a large, big,

221

extensive comments and opinions and

how we, how I did it and how we did it.

222

In that time we were, I was

working, preparing with one of,

223

other technical lead and we.

224

Go offline after that meeting.

225

And we have, a discussion with most

of the people that have commented that

226

have, some, opinion about, for us,

most of the opinions were in that time

227

weren't, like, agree with our decision.

228

And that was also what's kind of, make me.

229

Nervous in that time because you

are making the first decision

230

and a lot of people are okay, but

this is maybe not the right way.

231

You should do this and you should do this.

232

And we go like offline to each of them

and ask them like, and consider them.

233

Like I remember we take notes and we were

making a big moral board in that time.

234

Consider each of opinions and see,

okay, what should I do and what.

235

Should we consider this or not?

236

In that time, maybe some of the comments

or some of the opinions were because

237

people were not quite aware with the

product we are doing and everything,

238

but all of them were beneficial for us.

239

It's really good if you

have, comments and, opinion.

240

They are always welcome, even if

they're not, especially if people are

241

not agree with your opinion because

you might have missing something.

242

You might have not.

243

Not aware of something, uh, not, not have

that skills to make the right decision.

244

So it's, for me, it's really important

people to be encouraging to give

245

their decision, to give their advices.

246

how we continue, we ask people to

write them on the ideas itself.

247

We always asking them to write them on the

ads itself, so we have a track of them.

248

Kenny (Baas) Schwegler: So can I ask,

because you're saying opinions, right?

249

And from an advice process, you want to

have advice, instead of opinions, right?

250

So I'm also maybe curious

about Pete who is here, you, I

251

hear you're facilitating this.

252

And maybe both of you, it doesn't

matter who answered this, but how do

253

you deal with the fact that people are

coming in and I, oh, I'm an engineer.

254

I did it myself.

255

Strong opinions, weak,

256

Pete Hunter: Yeah,

257

Kenny (Baas) Schwegler: it again?

258

I, I, I don't remember the statement

anymore, but strong opinion.

259

Pete Hunter: strong opinions, weekly held.

260

Kenny (Baas) Schwegler: so from

both of your perspective, one

261

from a facilitator, but one from

actually, Elena, you were part of

262

the, you, you're coming with advice.

263

So maybe start with Pete and then

I'm curious about your, site, Elena.

264

How do you, how did you deal with that?

265

Pete Hunter: So I guess from my

perspective, it was, quite difficult

266

'cause you're trying to play Switzerland

in the advice process, right?

267

So you're trying to be impartial, you

are trying to facilitate conversation.

268

some of that conversation

will be opinion, right?

269

Because everyone does have an opinion,

and sometimes they're giving their

270

opinion with the best intention, and

sometimes they just wanna be heard.

271

so I think making sure there's space and

time for everyone to share their opinion.

272

But also kind of, I guess from

my perspective, I would, if, if

273

they're just forcing opinion.

274

would then go into the

role of asking questions.

275

Right.

276

So you start to unravel some of

their thinking and sometimes it is

277

just purely opinion, but sometimes

there's something behind it.

278

There's a reason behind it, and I

think that's why it's never worth,

279

even if you think it's an opinion,

because you've formed in a, a

280

kind of a, a view of that person.

281

it's worth digging into it, because

there will be something that they

282

have thought of that they think

impacts the decision you're making.

283

So even if you see them as opinions.

284

It's still useful information.

285

It still validates your decision.

286

It still kicks the tires on your decision.

287

Um, so it try not to, and I think

Eleanor was very good at this, even if

288

their opinions, you still validate them.

289

So yeah, that's,

290

Kenny (Baas) Schwegler:

I can guess, right?

291

People how to react can

also be like a gatekeeper.

292

Can imagine, right?

293

Because some people just don't

know how to express themself.

294

Maybe like me, English is not

my first, my first, language.

295

So expressing myself can

be maybe harder even.

296

so yeah.

297

Elena, how, how is that for you?

298

Because I can imagine, right,

dealing with criticism, because this

299

is what I'm seeing a lot, right?

300

You're, you're asking for advice

and you get harsh criticism.

301

How did you, how did you go with that?

302

Very curious.

303

Elena Stojmilova: I was used to opinion,

for the first a DR like, on purpose

304

because I must say, like most at that time

when we first started with architecture

305

advisory process and everything, most

of, our comments were, strong opinion.

306

We have a lot of people that have

like a strong, strong opinion.

307

We should do this and that

is the best way to do it.

308

And that's why we, uh, have at the

beginning how, at the beginning I

309

said, I have said like, it makes me

a lot of nervous because you are at.

310

First time making the decision.

311

You have a lot of people with different

and strong opinions, but how, we did that.

312

Like we have a conversation

with each of them.

313

We just appreciate their opinions and,

and, go in more detail because you.

314

It was necessary for me, to be more

sure that, we are not missing anything

315

and that we consider each of the,

each, each thinking in our decision.

316

I will say like some people, yeah,

it's sometimes it's part of the nature.

317

Some people will have

still strong opinion.

318

It's really important to

appreciate everyone opinion.

319

For me, I personally really like to hear.

320

More comments on my area.

321

Even if that is opinion not

advisors, I would like to hear them.

322

I'll say now, when we are in the

practice of the architecture advisory

323

Forum, we kind of used of how we

are adding the opinions Now most

324

of I, I can say that now we are in.

325

Better place comparing with that, uh,

uh, uh, four years, four years ago.

326

So it's not, yeah, there are still,

people with strong opinions in everything.

327

But most of the discussion on

our, on our arcs, uh, advice

328

advices, and we are mostly there,

is I am mostly receiving advices.

329

Should, can you try this

or should you try this?

330

If you consider this not just

strong opinion, this is the

331

best way how you should do it.

332

It'll need some time.

333

You kind of to, to, to be me as a person

who is writing the idea to, uh, to

334

kind of adjust on receiving all that,

opinions and advice and, consider them

335

in your decision making process, but

also to people who are coming to the

336

architecture advice to be prepared to see,

337

different opinion, different

decisions, and to cons to go

338

through them and consider them.

339

So also it'll need time for the

audience that is on Architecture

340

Advisory Forum to kind of adjust

and go with the, advice process.

341

Andrea Magnorsky: Seems like your org has

evolved quite a bit from that first, the

342

first time I, I'm kind of curious about,

you know, you kind of say, oh, these

343

days it's mostly advice, and was there

any particular one of these decisions,

344

like in between four years ago and

now, was there anything that was like.

345

I wish we would've done some tweaks.

346

I am sure the process changed

slightly from there and I, is there

347

anything there that is salient?

348

That's like a, I wish I would've

known that, about the advice form

349

itself and how to facilitate it.

350

'cause I think that seems non-trivial

and that you would learn a lot

351

from, from an experience like yours.

352

Pete Hunter: I mean, I mean, one

of the things that I've, I've kind

353

of reflected on, we went in from

a cold start, so straight into

354

an ice bath and figure it out.

355

I do think the, from a, architecture team,

who were present on that, first a f and

356

afterwards, it's quite a shock to them.

357

So I think they need a bit of a

longer lead into it than actually

358

the tech leads and the teams.

359

I think, communication skills

are really, really important

360

for, for that type of forum.

361

Also their role was changing, right?

362

So their role, the role of architects is

now not to hand down on a stone tablet.

363

The, you know, the architecture.

364

It's to, facilitate and share

their knowledge and help

365

people make better decisions.

366

But that's not, you know, if you've

not been doing that for the last.

367

4, 5, 10, 20 years, that's

not gonna come naturally.

368

So I think the lead up into that

for architects probably needs to

369

be longer than it does for the

teams that are actually making the

370

decisions, they'll run into it.

371

Um, but that, that, that brings

other problems that it's so.

372

Kenny (Baas) Schwegler: It's sort of

like architects have so much learned

373

how to do architecture in one way.

374

I remember.

375

To, learning the piano one time, like

I did it a year, but they say it's very

376

important to go to the teacher because

the moment you put your fingers in a

377

different way than it's useful, then

unlearning that takes a lot of time.

378

So I guess for a lot of architects,

what you're saying is Yeah,

379

Pete Hunter: Okay.

380

Kenny (Baas) Schwegler: learned one

way and now they need to unlearn that

381

way and learn it in different way.

382

While maybe tech leads and

people in teams, they are not

383

necessarily already learned in how.

384

that architecture works.

385

So for them, it's easier to learn

because they don't need to under.

386

Pete Hunter: Yeah, exactly.

387

Yeah.

388

It's, they're moving forward

outside of their comfort zone.

389

So yeah, it's all new to them with

architects, it is, you know, it,

390

it is about that, that they've

previously been on broadcast and

391

the deems have been on receive.

392

That broadcast now starts to

become a two-way conversation and

393

potentially a conversation that will

go in a different direction than

394

they would hope so.

395

They have to, Adjust their

muscle memory, I guess.

396

Andrea Magnorsky: Thank you very much

for, bringing us to this amazing story.

397

It's really nice to hear first

accounts and, yeah, thank

398

you for being here with us.

399

Kenny, Elena, Peter, see you next time.

400

Kenny (Baas) Schwegler: you next time.