Dominik Birkmeir & Andrea Eggert - Less tech, more value tacking digital waste with sustainble produ

This talk focuses on the concept of reducing digital waste in software engineering to enhance the value of products. The speakers, Dominic and Andrea from Myan Wolf, highlight that a significant portion of software never goes live or fails to meet expectations, indicating a hidden crisis in the industry. They introduce the idea of 'digital waste,' which includes unused features and complex designs that do not serve users effectively. The discussion touches on the importance of considering environmental, social, and governance (ESG) costs in software development and using methodologies like life cycle assessment to create more sustainable digital products. They suggest practical strategies to minimize digital waste, such as avoiding unnecessary features, being mindful of server usage, and optimizing design processes. The talk encourages developers and product managers to prioritize value over technologies like AI by ensuring that every feature and product iteration serves a genuine user need.

So, hello everyone. Um, it's kind of a weird situation that we're talking and don't see if you really hear me. So, give me a sign. Perfect. So, um, happy that you made it to our talk with the title less tech more value, which is certainly quite an out of the ordinary talk title for a tech conference, but I hope you bear with us and see what we

mean with it. by the end of this talk. We that's me Dominic, my colleague Andrea. Um I'm from Myan Wolf. My Wolf is a German software engineering company. So we are about a thousand uh developers developing custom um software in projects for large clients certainly in the automotive industry for example BMW but also quite another um area a large area um I myself I'm called head of

design and engineering at mywolf and Andrea >> hi my name is Andrea I'm head of design of new spectrum which is the design and innovation studio within my bomb wolf. I'm very passionate about creating our digitalized futures that we are actually craving to live in. So everything about innovation and what's up next. >> So to make that all more fun for you, we added two of the

presenters. So whenever one of us is talking too slow, we can just skip forward. Um but I hope we don't need it. Um, what if 22% of all buildings just collapse? Or 22% of laptops are crashing in an important presentation. I hope not this one. Or 22% of all cars are just dying down and are unable to start or 22% of all medicines not working. I mean,

the outcry would be big. You know, we would see a lot of stuff going on in the news and everyone talking about it and you would see everywhere that there are things lying around which just don't work. Strangely, with software engineering, we think, well, that's fine. 22% of our programs never go live, never work, but hey, that's fine. And why is that? We we really don't know

it. I mean, if you look at all the statistics, and this is from the Standish Group uh report, you can select whatever you want. Over the years, it's always the same. Roughly 22% of all software never goes live and is just not fulfilling the promises that were made initially. And another 49% are usually challenged. They not delivering what you promised it to be in the beginning. Why

do we accept that in software development? I have an idea and you can see it behind or better said you cannot see it because you just can't see what we are producing in software development. You can't see our digital waste that is lying around. It's no problem to see all the cars standing in the streets not working, the buildings breaking down. You can see that and you

would have to do things to get rid of all that. But with software, you don't see what things are not running on the servers anymore and all What do we mean by digital waste? We mean products. We mean features that are unused. I don't know who of you has ever developed a feature that went live but no user is actually clicking on that damn button because they're

just not wanting to use it or it's just too over complex which we are kind of um which kind of happens often that we're building software that's just too complex for our users and now with AI I mean everyone talking about AI it's like this this uh sledgehammer in your hand and everything is certainly a nail and you have to throw AI on it because we just

need to do it in our days. I don't think that's the best way and that's the way we produce a lot of digital ways. digital ways being the unused features, the unused things. But you can also look a little bit deeper not just on the tech side itself but also on the side when you look at uh socioethical aspects of software with the dark patterns we are

developing every day and we we are intentionally developing those dark patterns like doom scrolling. We want to have the user scrolling, scrolling, scrolling just because it generates revenue. Um, but that's also kind of a digital waste that we are generating or social scoring uh to name another one. So my message is avoid digital waste in your products and as Andrea is developing those products in the beginning,

it's your task Andrea. >> Cool. So thank you for handing over this uh small task. I mean, it's not like that product managers or designers just have nothing to do and why not getting a new task on our desk. When you think back to 10 years, uh you could be successful as a product manager if you just make your responsive because then it's available on all those

devices. 10 years later now you are asked to um make all the cultural nuances being represented by the AI in your product and actually please do it CO2 negative if you can as well and for sure reduce digital waste. So I thought well actually in my private life I'm not too bad and doing some good decisions on that topic. So I take a good challenge and well

it's actually also not that hard because the one thing how to tackle that very nicely is we just don't build products. That's certainly a nice idea. Um a product we never build it doesn't generate digital waste. So that's good. that was not quite what I was expecting from you. So really, we need to do our products because we want to go live at some point. But what

I want you is to consider what you're doing. Consider the project costs. Obviously, everyone is considering the project costs like what it costs to build the software, to run the software, to maintain the software. Certainly, everyone just does that. But I also wanted to consider that you take a look at the ESG costs. So the environmental, social and governance costs that has come with your software being

it the resources on the service the consumption of energy of water. I mean AI is consuming water more and more and drinking water not like the the dark water that you have in some um rivers and you're producing a lot of CO2 and unused hardware and all that. Consider that please and make sure you see what is the value of the software in the business value, the

customer value and make sure that the value always outweighs the cost. Well, okay. Luckily, this is not a new topic. It had been in physical product creation all around and there's a model that we could just follow along and let's see how we can adapt it to digital material. So it's a life cycle assessment. You might have heard about it. It starts with the natural resources you

are extracting from our mother earth to actually build those products. Then you have a look at how you could make the production more sustainable. Also, the packaging plays a big role there. Make sure you can ship more products with the same volume, for example. And then, of course, use a maintenance like the fair phone where you can just swap out parts and make sure you can repair

it. And then you can take care of the disposal, reduce as much as you can of waste treatment, and possibly recycle as much as possible. But what would that mean for the digital material? Well, so I'm not here to preach like the one and only way to be more sustainable and if you don't, you suck. But I'm here to give you like a couple of thumbs up

rules that we just have in mind in our daily work that really helps us making steps in the right direction. So in my opinion, we would have to apply this pattern three times for digital one, the servers your stuff is running on. Two, the user devices, and three, the digital product itself. So for the servers, well, most of this is not in our hands, but to keep

it simple, the grass is always greener on the servers of the big cloud providers. Why is that so? they have um an intrinsic motivation to make as much efficient usage of their server racks as possible. So for serving costs on their side as well. So also when for a product management point of view as much as you can reduce your server costs you're also reducing your ecological

footprint. But what's for the user devices? So your smartphones, your laptops, whatever you have on your drawers or in usage, what do you think is better? Less or more? Actually, in that case, the more the better. Why is that? So two things. First, if you can make your application accessible from as many devices as possible, you can also reach more people. And having more people reached means

that you don't need several applications being out there in the world to do the same thing. And the second thing is please make sure to find out what your customer base is really also having devices and usage and keep supporting the older devices as long as you can because then they do not have to buy newer devices which of course has its ecological footprint. But what about

the digital product itself? So here's um a thing a couple of methods. It's a huge bunch you could use that I really like to use because they just fit in all the task we are already doing because it's for me now no customer actually paid me more to be more sustainable there. So I really like to have it efficiently implemented. So one method I really like is

the futures wheel. It is a method of design futuring which helps you to map out um in the middle of your map an event, your product or a trend and then you brainstorm in a group of people the first level direct impact of that event and then even go to the second level of the impact the impact of the impact there and then after all you have

a whole map you can take as a discussion for your product idea in the very um first stage of your product design and see what you like about it, what not and you have a direction of the ideal world where you want to head to. And the second method I really like is later on who of you had done a design sprint before know how it works.

Okay, so a design sprint starts on a Monday. you come to a group together, you uh tackle a problem and during one week you build an actual pro um prototype to solve that customer problem. At the end on Friday you would actually have real people ideally potential customers try out the prototype and let you know what you think. And there you can implement non-human personas. Just think

of it. You can find them online. There are templates in Figma and Miru where everywhere you can use them. You can make sure that they are updated by maybe doing a little bit research of yourself and then having with a nice prompt for example chatbt or another AI tool roast your idea from that perspective. So here are the two things for discovery. And then when it comes

to delivery, two more things. The one is I think that's nothing new and you're probably more the expert on this than I am as a designer. You reduce your bundle size. Meaning for example, somebody um enters a website and the website would ultimately load all the data that could be displayed there. It's much better ecological wise to only load the ones that users or visitors actually need.

For example, use pagionators or only if they interact with the um data actually load the data. Um and the next one is watch out for the zombies in your project. Meaning if you have for example less than um 0.5 uh percentage usage of certain features, it's most likely that you could just delete them and get them out of the way. That means you don't have to run

them all the time. You don't have to update them all all the time and you could just get rid of them or maybe um just really find out what you need to improve. So because it's a core feature or it could be and then find out how to improve that feature. >> So all the methods and all the theories is always fine. But let me guide you

to two um stories um of products we actually developed or did not develop during the last couple of uh months and years. One is about a home appliance manufacturer. So, a manufacturer of ovens and washing machines and all that. And they had an amazing idea. They had this great idea that you could use Alexa to control your washing machine. Isn't that great? I mean, your washing machine,

you can talk to it now. You can tell it that it needs to wash your laundry. I think it's a good idea. But thinking about it twice is well, but I'm I'm kneing in front of this washing machine because I have to because it doesn't go in my room and collects all the dirty clothes from my floor. So, I have to go there and put it in.

And guess what? It's so much easier to just push that one button or turn the wheel a little bit instead of using Alexa to do anything. So this is really a use case that does not work at all. But they developed it. They used that and that was the first thing they developed for Alexa. The first Alexa skill was control your washing machine. Amazing. like if you

work into it and using all those methodologies that Andrea just talked about you can find good use cases because for example if you don't look at the washing machine but the oven sometimes you say why the oven I also have to go there and put in my my cookies that I'm just preparing. Yeah, sure. But while you're netting the dough still and working hands sturdy in it,

telling your oven to preheat for a certain temperature, that's quite a nice feature as long as you don't have anything in there that shouldn't be there. But that is quite a nice feature. And even maybe telling it the recipe and it um is smart enough to find the correct temperature on its own. Those are things that really generate some value. But right now, as I said before,

all our customers, they come to us and they want to do something with AI just because you have to. I don't know if you had any customer in the last couple of uh months or you it yourself at the company haven't had this idea but everyone wants to do something with AI right now and what is the result that most of the time turns out as something

with AI is something like that some chatbot where you can type in text and you get back back text isn't that great I mean we had that couple couple of years ago like 30 years where you had this mainframe things you were typing in text getting bad text we had the same concept already but now we're using it again and guess what most of those I I

don't have anything against genai they have really a very good value but most of the chat bots and all that that is developed during the last couple of years people just don't use it they use it once they use it twice but then they at some point they get bored and say this doesn't really generate any value for me. Um, I like this citation which uh was

in a newspaper I recently read. It says the same applies to AI as to drunk texting. Just because it works doesn't mean it's a good idea. I think that describes the scenario quite well. And now one of our customers, I can't tell you the name, but we have like transferred this exact um example to one that I can show you now is we call a sports companion.

Okay. First clue is a chat. Yeah. Do you have a chat where your sports companion you're going into a new city like because you're on a business trip and you want to find a yoga studio for your evening workout. quite a nice idea. You used this chat. Um it is already connected to your um to to the studios that you're allowed to visit and you can find

something. But instead of just putting down text and getting text in text out, we enhanced it a little bit by well giving back some um forms already being able to just pick a date and a time and all that, which was quite a nice thing to improve the simple text chat. But the real thing went on when we integrated it in a completely responsive web page so

that if you use the chat and you're doing something in the chat, it automatically updates the map next to it. It automatically shows you um this different studios and options that you're having. And as soon as you click on something or interact with it, also the chat goes one step further. And that was a real new experience and people love to use that actually. So coming back

to our initial claim, you always have to have a look at the costs of your product and the value of your product. And you need to make sure cost. And with that I mean really looking into using less resources, using less energy and all that and increasing the value as far as possible. I wasn't sure if the next slide is mine. It is still mine. um the

whole methodology behind that. I really would have loved to show that to you right now, but we don't have that much time because it will certainly take more than just 25 minutes for a talk. So I would invite you to well get one of our books that we have um at our booth and also you could access the white paper there or sign up for the white

paper with design for AI. back to you and >> so I hope we could motivate you today to channel your inner Aria a little bit and maybe come over to our booth over there. It's actually really over there. You can see the posters and we can inspire each other a little bit into the more sustainable digital design So, thank you [Applause]