[00:00:00] [music]
[00:00:04] >> Welcome back to another episode of Tech
[00:00:06] on the Hinch where tech gets human
[00:00:07] powered by Code District. I’m your host
[00:00:09] Rabia Javed and my guest today is Arnie
[00:00:11] Kittler who is both a long-time product
[00:00:13] leader and the co-founder of Product at
[00:00:15] Heart, one of Europe’s leading product
[00:00:17] conferences. He trained media economics
[00:00:19] then moved into product leadership at
[00:00:21] Xing and now leads product
[00:00:22] transformations for clients across
[00:00:24] e-mobility, SaaS, and publishing. He’s
[00:00:26] also a co-founder of Hey Clarity. Today
[00:00:29] we are talking about what AI still can’t
[00:00:31] do in product. Arnie, welcome to the
[00:00:32] show.
[00:00:33] >> Hello, hey, and thanks for having me
[00:00:34] today. I’m really glad to be on the
[00:00:36] podcast. Look forward to your
[00:00:37] conversation, Rabia.
[00:00:38] >> Arnie, you’ve done a fair number of
[00:00:40] things before getting into product. You
[00:00:42] did record label, fanzine, actual
[00:00:45] touring, musician stuff. What’s the one
[00:00:47] skill from that world that still helps
[00:00:49] you now?
[00:00:50] >> One thing that I learned very early on
[00:00:52] in in these experiments, maybe you could
[00:00:54] say, I mean these were not actual
[00:00:56] commercial endeavors but more niche
[00:00:58] hobby things, was that I’m usually
[00:01:00] better when complementing my views with
[00:01:03] others, so when collaborating with
[00:01:05] others. That’s better than if I’m just
[00:01:06] pushing through one thing by my own. And
[00:01:08] this is something that, you know, I’m
[00:01:10] I’m carrying with me throughout my whole
[00:01:12] professional life.
[00:01:12] >> You know, when you are worried about AI
[00:01:14] hallucinating and, you know, slipping
[00:01:16] into real product work, is there an
[00:01:18] actual framework or checklist your teams
[00:01:20] run to catch it?
[00:01:21] >> The best thing like when working with
[00:01:23] humans is to to question the responses
[00:01:26] you get and say, “How did you come to
[00:01:28] this conclusion?” or “Can you tell me
[00:01:30] where you found a certain piece of
[00:01:32] data?” I mean this is not really a
[00:01:33] framework, this is more just common
[00:01:35] sense of how you interact with systems
[00:01:38] or with people for high-stake matters,
[00:01:40] but that’s what I would usually do.
[00:01:41] >> Makes a lot of sense. [clears throat]
[00:01:42] So, Arnie, when we talk about, you know,
[00:01:44] AI across the organization, we see that
[00:01:46] AI productivity gains get hyped
[00:01:49] constantly. How do you actually see that
[00:01:51] rolling out in the product?
[00:01:53] >> One disclaimer I want to make, I’m Uh,
[00:01:55] speaking here from a European
[00:01:57] perspective and as somebody who has
[00:02:00] who’s mostly working with scale-ups and
[00:02:02] enterprises. So, what I’m saying may not
[00:02:04] reflect the reality of North American
[00:02:07] early-stage startups, for instance. I
[00:02:08] know that the dynamics may be a bit more
[00:02:10] energetic, let’s maybe say it like this,
[00:02:12] or a bit a bit further uh in the
[00:02:14] adoption of AI. But, what I’m seeing
[00:02:15] right now in in my product work across
[00:02:17] different organizations is that still
[00:02:19] the focus very strongly is on the
[00:02:22] individual enhancement. Individuals
[00:02:24] trying to complete tasks better or
[00:02:27] create artifacts in a quality or in a
[00:02:29] way that they weren’t able to do so.
[00:02:31] And, to be honest, this is this is good
[00:02:33] and I’m seeing positive effects of that,
[00:02:35] but this is not and I I I don’t think
[00:02:38] that this will ever lead to 10x
[00:02:40] productivity gains that some, you know,
[00:02:43] people in upper management positions may
[00:02:45] be hopeful when when their organization
[00:02:47] adopts AI. And, I think for more
[00:02:49] impactful productivity gains, I think we
[00:02:51] need to think beyond the individual
[00:02:53] enhancement and rather think about how a
[00:02:56] software development team can
[00:02:57] collaboratively use AI. And, this is
[00:03:00] something that is still relatively new
[00:03:03] to most of the European companies that I
[00:03:05] work with. Some are now establishing
[00:03:07] their own team OS’s as a as a central
[00:03:10] compounding reference in their
[00:03:12] collaboration. And, I think these are
[00:03:13] really promising developments, but I
[00:03:16] think even with that 10x productivity
[00:03:18] gain expectations are still rather
[00:03:21] far-fetched, especially also when you
[00:03:23] when you are not working in greenfield
[00:03:26] environments or when you’re working in
[00:03:28] complex organizations, because in the
[00:03:30] end, I mean, we still need to make
[00:03:32] conscious decisions throughout the
[00:03:34] process. And, even though the steps in
[00:03:36] between may speed up quite dramatically,
[00:03:39] if the organization is not yet ready to
[00:03:42] also 10x their decision velocity, then
[00:03:45] we cannot really uh have 10x
[00:03:48] productivity gains. We will, you know,
[00:03:49] be faster in the in-between steps, but
[00:03:51] then run into decision bottlenecks or
[00:03:54] other, you know, things where the
[00:03:55] somewhat slower adopting culture of the
[00:03:59] organization needs to catch up with the
[00:04:01] immediate productivity gains we’re
[00:04:03] seeing in how we apply AI for certain
[00:04:05] tasks.
[00:04:05] >> Across your clients at the moment, did
[00:04:08] adding AI tools actually simplify
[00:04:10] anyone’s stack, or is it just one more
[00:04:12] system someone has to manage now?
[00:04:15] >> right now what I’m seeing is more when
[00:04:17] I’m speaking specifically about the
[00:04:18] product development cycle, I think that
[00:04:20] AI tools such as Claude or Claude or
[00:04:23] Blod were added to the mix and are used
[00:04:26] to, you know, operate existing solutions
[00:04:29] like Jira or Figma and so on, or connect
[00:04:32] them in in in different ways that we
[00:04:34] couldn’t before and and, you know,
[00:04:35] smooth in processes and take on certain
[00:04:37] things uh of the responsibility, but I
[00:04:40] haven’t seen them like fully replace any
[00:04:42] of the established solutions yet. And uh
[00:04:45] but I mean, I agree with the notion of
[00:04:47] your question. I think everybody would
[00:04:48] like more simplicity, but like right
[00:04:51] now, for instance, I remember when when
[00:04:53] Claude design was introduced, of course,
[00:04:56] my teams were hopeful that uh you know,
[00:04:59] at some point they would be able to
[00:05:00] replace Figma, for instance. And like
[00:05:02] right now, I mean, we’ve tried it and
[00:05:04] that that that that doesn’t seem to work
[00:05:06] out yet. I mean, you know, things evolve
[00:05:08] and and at some point we may be able to,
[00:05:11] you know, the the AI tools may replace
[00:05:13] some of the established players, but on
[00:05:15] the other hand, we also have to
[00:05:16] acknowledge that companies like
[00:05:18] Atlassian or Figma. There’s a reason why
[00:05:21] they are where they are. So so I mean,
[00:05:23] they will probably also do their best to
[00:05:25] to keep their products relevant by by
[00:05:27] integrating AI into them. So let’s let’s
[00:05:29] see what happens.
[00:05:29] >> Yeah, no, that makes a lot of sense. And
[00:05:31] you know, even till today and to this
[00:05:32] point, they are trying to catch up, you
[00:05:34] know, with the advancements of AI, given
[00:05:36] that, you know, they’ve had to see some
[00:05:38] major losses lately. And hence, you
[00:05:41] know, Figma came up with all the recent
[00:05:43] integrations, you know, across the AI
[00:05:45] tools. And um so did it last year. But
[00:05:48] now, you know, people used to manage
[00:05:50] their entire workflows primarily or
[00:05:52] priorly on that, they are now making
[00:05:55] their own workflows. So, to some extent,
[00:05:57] there yes is a lot of effect, but in
[00:05:59] order for these particular tools to stay
[00:06:01] relevant, they um they got to, you know,
[00:06:04] keep catching up.
[00:06:05] >> it’s much different like, I mean, maybe
[00:06:06] just as as a side note, I mean, what I’m
[00:06:07] seeing in the organizations that I work
[00:06:09] with is that when it comes to, say,
[00:06:11] internal tooling for more general
[00:06:13] managerial matters like, I don’t know,
[00:06:15] travel planning and so on. I mean,
[00:06:17] that’s an area where I see, you know,
[00:06:19] many more in-house tools replacing the
[00:06:22] likes of Monday and so on.
[00:06:24] Company-specific white-coded solutions
[00:06:26] or, well, maybe not even white-coded,
[00:06:28] but like AI-built uh solutions for
[00:06:30] simple tasks and and I mean, this is
[00:06:32] this is something that I’m seeing being
[00:06:34] rolled out more intensively than than in
[00:06:37] the in the product development cycle
[00:06:38] right now.
[00:06:39] >> Martin, you know, how do you actually
[00:06:40] feed confidential data, you know, um of
[00:06:43] your clients into AI tools without
[00:06:45] risking that sort of confidentiality? Is
[00:06:47] there a real policy or mechanism behind
[00:06:49] it that you follow or your company
[00:06:51] follows?
[00:06:51] >> Well, I mean, I think my own policy is
[00:06:54] that I won’t violate my clients’
[00:06:56] uh confidential data. Uh so, I won’t
[00:06:59] feed it into the system. I mean, the
[00:07:00] thing is that, you know, I act in
[00:07:01] different roles. Sometimes I act as a as
[00:07:03] an advisor or as a as a coach where I
[00:07:05] don’t have access to the client’s
[00:07:07] systems. In that case, I will be very
[00:07:08] careful with what how I use the
[00:07:10] information. In other cases, well, like
[00:07:11] right now, I’m also doing interim head
[00:07:13] of product role where I’m embedded in
[00:07:15] the organization. And in those cases, I
[00:07:17] have access to the organization’s
[00:07:20] shielded AI environments where, you
[00:07:23] know, that that are safe to to also work
[00:07:25] with the company’s data. That that, of
[00:07:26] course, I will do. I will be careful.
[00:07:28] And and if I’m interacting outside of of
[00:07:31] the client’s uh safe environments with
[00:07:33] with AI, I will, for instance, always
[00:07:35] use code names instead of the real
[00:07:37] client or product names to make sure
[00:07:39] things are safe.
[00:07:40] >> And you know, are there is there
[00:07:41] anything that you tell your teams about,
[00:07:43] you know, that how should they be
[00:07:44] careful while adding data on AI or using
[00:07:48] AI?
[00:07:49] >> Yeah, I mean, to be honest, in the
[00:07:50] organizations where I’m more embedded
[00:07:52] and have actual responsibility for
[00:07:55] uh for teams, you know, they have
[00:07:56] general policies in mind. So, I’m I’m
[00:07:58] not the main advocate for it. I I don’t
[00:08:00] need to be because because there
[00:08:02] thankfully are systems in place for, you
[00:08:05] know, how they can
[00:08:06] interact with the foundational models.
[00:08:08] There are also systems in place where
[00:08:09] they can deploy stuff and so on. So, so
[00:08:12] thankfully I can I can build on this.
[00:08:14] This may be a German thing, you know,
[00:08:15] German Germans pay attention to those uh
[00:08:17] things. So, so many German companies
[00:08:19] have been quick to to introduce
[00:08:21] environments that at least seem to be
[00:08:24] safe for
[00:08:25] uh for the use of AI.
[00:08:26] >> And hence the, you know, EU AI Act as
[00:08:29] well that, you know, um got out way
[00:08:31] sooner than the rest of the countries,
[00:08:33] you know, were looking at it. So, you
[00:08:34] know, that makes a lot of sense. Adding
[00:08:36] an extension to your point. So, are they
[00:08:38] um you know, you’ve talked about being a
[00:08:40] translator. You know, you’re a board
[00:08:43] advisor to at the moment. So, you know,
[00:08:45] being the using CFO language for the
[00:08:47] CFO, customer success language for the
[00:08:49] CS. Can AI actually do that translation
[00:08:52] or does it come out sounding the same no
[00:08:54] matter who it is for?
[00:08:55] >> Well, I mean, you can use AI for that.
[00:08:57] But I mean, the main thing to
[00:08:59] acknowledge as a as a product person is
[00:09:01] that it is on us to make ourselves
[00:09:03] understood. So, so like what you will
[00:09:05] sometimes hear when you speak to product
[00:09:07] focuses is yeah, you know, the the sales
[00:09:09] folks didn’t get it again. And I mean,
[00:09:11] it it’s you may think so, but in the end
[00:09:13] still it’s your responsibility to make
[00:09:15] sure that your sales team is well
[00:09:17] enabled and gets what you’re doing and
[00:09:19] why you’re doing it so that they can
[00:09:21] represent your product to the outside
[00:09:23] world and you can together be
[00:09:24] commercially successful. So, this is
[00:09:26] something that I I speak about a lot.
[00:09:28] Acknowledging that importance of making
[00:09:30] oneself understood. And as you said,
[00:09:31] like this means you speak to a CFO about
[00:09:34] money and you speak to customer success
[00:09:36] folks about, I don’t know, customer
[00:09:37] retention for instance, you can use AI
[00:09:39] to to help you with those translations
[00:09:42] and I know that for instance some people
[00:09:44] that I work with that have difficulties
[00:09:46] justing their messaging for different
[00:09:48] audiences that they use AI as a sparring
[00:09:50] partner and say, “Hey, okay, this is
[00:09:52] what I want to say. Do you think this
[00:09:53] will come across well when when I’m
[00:09:57] talking to a given group or not?” So,
[00:09:59] yeah, as a sparring partner, yes, and
[00:10:01] then it can be helpful. Maybe one one
[00:10:03] small plug here if I mean in general if
[00:10:06] product folks are listening who want to
[00:10:07] get better at well, getting their points
[00:10:10] across to top management, I can really
[00:10:11] recommend Mick Rich Mironov’s recent
[00:10:14] book
[00:10:15] called Money Stories in which he talks
[00:10:17] exactly about this like how do you, you
[00:10:19] know, put all the product thinking that
[00:10:20] you have into numbers that will resonate
[00:10:23] with the top management and will go,
[00:10:24] “Aha, okay, that’s that’s why we should
[00:10:26] invest in your team and so on.” So,
[00:10:29] really good book, Money Stories.
[00:10:30] >> I’m definitely going to have a look at
[00:10:31] that. Okay, so Arnie, you know, once you
[00:10:33] are known as the AI fluent one on the
[00:10:36] team, are you still allowed to say,
[00:10:38] “Hey, I don’t get this?” Or does that
[00:10:40] reputation then mean you are the one
[00:10:42] who’s not allowed to be confused
[00:10:44] anymore? You know it all.
[00:10:45] >> I mean, I would say that right now I
[00:10:47] don’t I don’t think that anyone is
[00:10:49] currently done with learning. So, so I
[00:10:52] would say even the most AI fluent
[00:10:54] people, which maybe I’m not, still they
[00:10:56] will need to learn things every day and
[00:10:58] I think this is also part of being AI
[00:11:00] fluent meaning having found your ways of
[00:11:03] learning and and deciding which of all
[00:11:06] the noise and all the the new things
[00:11:07] that are happening out there you
[00:11:09] actually take into consideration and and
[00:11:11] learn about and which of the things you
[00:11:13] just say, “Okay, you know, you cannot
[00:11:14] focus on everything right now, right?”
[00:11:16] So, so I think building the skill of
[00:11:17] learning is super important right now.
[00:11:19] For me as a as a leader, I think it’s
[00:11:21] particularly important to also always
[00:11:24] acknowledge that you don’t know
[00:11:25] everything because if you don’t and if
[00:11:27] you pretend you know everything, I mean,
[00:11:29] first of all, like right now as I said,
[00:11:31] I don’t think there’s anybody in the
[00:11:33] world who knows who’s right about
[00:11:34] everything. And by acknowledging that
[00:11:36] you don’t know everything either, you
[00:11:38] also encourage, well, the rest of the
[00:11:40] organization to to constantly learning
[00:11:43] organization, and you encourage people
[00:11:45] to apply, I think what’s sometimes
[00:11:46] referred to as a growth mindset, you
[00:11:48] know, instead of saying we know it all,
[00:11:50] let’s all think about how we can further
[00:11:52] grow together, how we can learn
[00:11:53] together, and so on.
[00:11:54] >> So, you know, through Hey Clarity, you
[00:11:56] coach people to turn a vague half-formed
[00:11:58] complaint into a real problem worth
[00:12:00] solving. Is that actually teachable, or
[00:12:02] is it something people just develop over
[00:12:04] time?
[00:12:05] >> Well, I mean, as a when when coaching
[00:12:06] people, of course, the part of what you
[00:12:08] do is you help people develop that by
[00:12:10] pointing, well, opening sometimes also
[00:12:13] new perspectives for them. And something
[00:12:15] that I’m noticing when I’m coaching
[00:12:17] product leaders is that what they
[00:12:18] struggle with often are not product
[00:12:20] management techniques also. They they
[00:12:22] they know this. They’ve read the books.
[00:12:24] They’ve they’ve done the jobs. Often,
[00:12:26] it’s about some sort of stakeholder
[00:12:28] management or expectation management.
[00:12:30] And so, in many cases, it is all like
[00:12:32] also again about this question of making
[00:12:34] yourself understood. And for instance,
[00:12:37] thinking about bringing the important
[00:12:39] initiative you want to drive across in
[00:12:42] different ways to different members of
[00:12:44] of the C-suite as an example. So, yes, I
[00:12:46] think if this is somewhat, I mean, not
[00:12:48] teachable, but you can help people
[00:12:51] develop their own feelers and then build
[00:12:53] build on that with their own in their
[00:12:56] own personal way.
[00:12:57] >> So, you know, we see R&A that new grad
[00:13:00] hiring at Big Tech has reportedly gone
[00:13:03] from 15%
[00:13:05] to 7% in just 3 years. If that entry
[00:13:07] door keeps closing, where does the next
[00:13:10] generation of product judgment come
[00:13:12] from?
[00:13:13] >> Yeah, I mean, this is a question that is
[00:13:14] on my mind a lot. Also from a personal
[00:13:17] reasons, like I’m the father of a
[00:13:18] 21-year-old daughter who studies digital
[00:13:20] management. And so, so she is in that
[00:13:23] position. And indeed, it is true, like
[00:13:26] right now, it’s very difficult to land
[00:13:28] entry jobs or even internships because
[00:13:30] even for internships, you’re expected to
[00:13:32] already have previous experience, but
[00:13:34] you have to start somewhere, right? And
[00:13:36] I mean, like luckily, she’s now found a
[00:13:38] found a great company that that she’s
[00:13:39] doing the internship with, but it was it
[00:13:41] was not so easy, to be honest. And the
[00:13:43] other thing that I’m I keep thinking
[00:13:45] about is that I mean, the the next
[00:13:47] generation of product people, how they
[00:13:48] will build product judgment or product
[00:13:51] sense will certainly be different from
[00:13:53] how my generation learned it. I mean, in
[00:13:55] in my case, I believe that a lot of that
[00:13:58] learning was about, you know, wrangling
[00:13:59] with the raw material, you know, sorting
[00:14:02] post-its from observations of of user
[00:14:04] interviews, really thinking about how to
[00:14:06] best structure a presentation, wrangling
[00:14:09] with the data to to to put it into a
[00:14:12] dashboard, and so on. And I mean,
[00:14:13] realistically, my daughter’s generation
[00:14:16] will not be doing those things manually
[00:14:17] anymore. And And neither would I, to be
[00:14:19] honest. I don’t know exactly how they
[00:14:22] will build product sense instead, but I
[00:14:25] must admit, I don’t I really don’t know
[00:14:26] how. And so so right now, I’m just, you
[00:14:29] know, I’m I’m usually a very optimistic
[00:14:30] person, so I I hope that this generation
[00:14:32] also will find their way. And I mean,
[00:14:34] one other thing where probably the more
[00:14:36] senior people that interact with them
[00:14:38] can help with is, you know, just to keep
[00:14:40] challenging them and pointing out areas
[00:14:44] where they may have, you know, just
[00:14:45] taken the easy route and the shortcut
[00:14:47] and and by making them aware that they
[00:14:49] haven’t really understood why they
[00:14:52] actually propose a certain product
[00:14:54] direction or so on. That way, helping
[00:14:56] them to always also reflect on what
[00:14:58] they’re doing.
[00:14:59] >> So, you know, Arnie, at one of your own
[00:15:01] leadership events this year, your
[00:15:03] organizers deliberately played down AI,
[00:15:06] betting that product leadership in the
[00:15:08] AI era is still just product leadership.
[00:15:10] Walk me through that.
[00:15:11] >> Play down AI. I mean, there’s no
[00:15:13] question that this is the most impactful
[00:15:15] change happening to product development
[00:15:18] that I’ve witnessed over the last 25
[00:15:20] years, you know, bigger than
[00:15:21] introduction of Agile or mobile or
[00:15:25] learning to work remotely and so on. I
[00:15:26] mean this affects everything. It affects
[00:15:28] our products and it affects how we work.
[00:15:30] But since we I mean at our leadership
[00:15:32] event we had we had 60 CPOs and other
[00:15:35] product leaders together and we were
[00:15:37] assuming that in the breaks and also
[00:15:40] some of the, you know, other more free
[00:15:42] conversation aspects of the event we had
[00:15:44] they would be talking about AI matters
[00:15:46] anyways. And instead what we wanted to
[00:15:48] focus on with the content and and the
[00:15:50] workshops that we brought to them were
[00:15:52] some other aspects that matter no matter
[00:15:54] if, you know, building products with AI
[00:15:56] or or not like, you know, understanding
[00:15:59] and clearly communicating why you’re
[00:16:00] doing something. That is important with
[00:16:03] or without AI. Or the importance for
[00:16:05] leaders to provide the necessary context
[00:16:08] so that everyone and everyone means
[00:16:09] humans and non-humans can do their their
[00:16:13] best job. Or like the importance of
[00:16:15] building trust with your team, with your
[00:16:16] stakeholders, with your customers, which
[00:16:18] is something that usually happens well,
[00:16:21] in virtual rooms or in actual rooms and
[00:16:23] and in, you know, between humans. And
[00:16:25] lastly also this the importance of
[00:16:27] bringing the organization along in
[00:16:28] whichever way you do it. I mean those
[00:16:30] are things that are important now, you
[00:16:32] know, they are not new. They they are
[00:16:33] just fundamental skills that that you
[00:16:36] need as a as a good product leader. And
[00:16:39] those were the things that we talked
[00:16:40] about and
[00:16:41] yeah, I think that also considered
[00:16:43] helpful by everybody who participated.
[00:16:45] >> If we conclude our entire conversation,
[00:16:47] what would you say that, you know, is
[00:16:49] still human and something that AI cannot
[00:16:52] perhaps do. You know, people are just
[00:16:54] live coding products and different
[00:16:56] applications and it’s all just super
[00:16:58] easy even if you’re not from coding
[00:17:00] background. But given that, you know,
[00:17:02] you’ve had so much experience in
[00:17:04] product, what would you say is still out
[00:17:06] there which is uniquely human and cannot
[00:17:09] be replicated by AI as of now?
[00:17:11] >> Yeah, I mean that’s I think that’s an
[00:17:13] important disclaimer to say as of now we
[00:17:14] we don’t all we all cannot know how
[00:17:16] things evolve, right? But like I’d like
[00:17:18] to think that everything well around the
[00:17:21] relationships between humans. I mean, in
[00:17:22] the end we are not building products for
[00:17:25] machines, we are building products for
[00:17:26] humans and and that comes, you know,
[00:17:28] with a human relationship to the users
[00:17:30] of our product and understanding those.
[00:17:32] I think, you know, AI can help put
[00:17:34] together the data, but still, you know,
[00:17:35] it’s it’s still a different matter if
[00:17:37] you are speaking to an actual client or
[00:17:38] an actual user. So, I think this is this
[00:17:40] is something that will still be
[00:17:42] important as well as within an
[00:17:43] organization, you know, creating trust,
[00:17:46] creating alignment, making sure that
[00:17:48] everybody’s pulling in the same
[00:17:49] direction. I think this requires some
[00:17:51] human empathy and and and skills that
[00:17:55] are right now difficult to to delegate
[00:17:58] to AI and I mean, I would also say when
[00:18:01] it comes to to creativity, that’s
[00:18:03] probably another area where where I’m
[00:18:05] happy that we have humans, that we also
[00:18:08] typically in cross-functional teams have
[00:18:09] humans with different backgrounds who,
[00:18:11] you know, bring in their their different
[00:18:13] perspectives. I think there is power in
[00:18:15] there that right now AI cannot replace
[00:18:17] yet and can only complement, but that’s
[00:18:20] a good combination in my view.
[00:18:21] >> Yeah, no, great answer. So, you know,
[00:18:23] before we conclude,
[00:18:25] two points. Arnie, what’s the hardest
[00:18:27] lesson you’ve learned building a
[00:18:28] product? Something you only understood
[00:18:30] after it went wrong and one advice for
[00:18:33] all the chief product leaders out there.
[00:18:36] >> I mean, let’s let’s come to the first
[00:18:38] difficult lesson. It was not actually
[00:18:40] one where I learned something new, but
[00:18:41] it confirmed something that I had
[00:18:43] forgotten. So, so, so basically, I was
[00:18:45] once tasked to build a new app for the
[00:18:48] company that I was working with and we
[00:18:50] had basically certain assumptions for
[00:18:53] how that app would, after we had built
[00:18:56] it, be distributed. So, basically, we
[00:18:57] relied on a certain user group to to
[00:18:59] say, “Hey, this is great. I will share
[00:19:01] this with my network and I will
[00:19:03] propagate that app.” And we assumed that
[00:19:05] that way, you know, our app would easily
[00:19:07] get the necessary traction for for it to
[00:19:10] to sense. And I did for various reasons,
[00:19:12] but but I have to admit I also
[00:19:14] overlooked it, even though I could have
[00:19:15] known better. I did not validate whether
[00:19:18] this assumption that this user group
[00:19:20] would actually, you know, help us
[00:19:21] promote the app and basically be the
[00:19:23] main growth driver for of the app, that
[00:19:25] would actually, you know, they they
[00:19:26] would be doing it. So, that reminded me
[00:19:28] that, you know, we really really need to
[00:19:30] be conscious about the assumptions that
[00:19:33] we have, and and before we invest a lot
[00:19:36] in extensive developments, make sure
[00:19:38] that the underlying assumptions as
[00:19:40] validated as they can be. I mean, it’s a
[00:19:42] dynamic environment, so you can still be
[00:19:44] proven wrong later, but in this case, I
[00:19:46] didn’t do my homework well enough, and
[00:19:48] since then, I’ve paid much more
[00:19:50] attention to getting that doing that
[00:19:51] better. Yeah, I mean, I think the the
[00:19:53] main advice I would have for product
[00:19:55] leaders out there is to still remind
[00:19:58] your teams that, despite all the
[00:20:01] fascination about building stuff, and
[00:20:04] everybody being a builder, and so on,
[00:20:06] you have to remind your people that, at
[00:20:07] the end, there are also still some core
[00:20:09] responsibilities that they carry as
[00:20:11] product managers, and it’s very easy
[00:20:14] right now to to be carried away by, you
[00:20:17] know, building prototypes just because
[00:20:18] we can. And I’m not saying that building
[00:20:20] prototypes is bad, but if, because of
[00:20:22] building prototypes, you neglect your
[00:20:25] stakeholder management, or your backlog
[00:20:26] is not in a good shape, or you forget
[00:20:29] about proper, you know, user research,
[00:20:31] then this is not helping matters, and
[00:20:33] and so, your primary job as a product
[00:20:35] person,
[00:20:36] they haven’t changed so much, and that’s
[00:20:38] remind the product leaders to teach
[00:20:40] their teams.
[00:20:40] >> Well, Arnie, thank you so much for such
[00:20:42] a grounded and honest conversation with
[00:20:44] us today.
[00:20:45] >> [music]
[00:20:45] >> You’re welcome.
[00:20:51] >> [music]
