We discuss the biggest wastes of money for devtool startups, plus how to think about design partners vs contractual commitments. Bonus topic: another boring launch.
October 23, 2024
-
21
mins
NOTES:
Ads are a waste for most early b2b companies: https://x.com/karrisaarinen/status/1846988080143913056
AWS builds Iceberg on S3 for Netflix: https://x.com/marcoslot/status/1844849407285407968
Warp's rebrand launch: https://x.com/warpdotdev/status/1847306687788589119
TRANSCRIPT:
Hank: But generally, unless you're running really good tests, I would say ads are just giving money, giving your VC money right back to Google.
Gonto: On top of that, there's a couple of things that are interesting to me. Like number one is most developers will not click ads because they have ad blocker or something like that.
Even now, a lot of browser were, not Chrome, because Chrome now disallowed a lot of the ad blockers. But if you use ARC, you use Brave, or some of the others. You'll get them mostly disabled, um, by default. So a lot of these ads are not even seen.
Hank: Gonto almost every client I have. I've told them there's one way they could waste a lot of money and flush it down the toilet. Do you know what it is?
Gonto: What is it?
Hank: It's ads. And this topic's brought up because we saw, uh, Carrie from linear mentioned that they only spent 30K on ads. It wasn't even proper ads. It was 30K on sponsorships of podcasts. I think one of them was Lenny's podcast. He tagged on there.
And I really liked that take. Um, ads are really difficult before you have product market fit for a ton of reasons. You don't know exactly who you're talking to, so it's really easy to just write really generic, basic copy that no one wants to click.
It's really easy when you don't have someone more specialized or experienced, which you don't at the start of a company, to under optimize and over generalize your keywords, um, or to just have bad images, etc, etc, etc. Plus, you don't know how to do the attribution or tracking or any of that. So you have no idea what your ROI is or if you're even getting anything from it.
Sounds like you feel the same, maybe there are other ways to waste money too.
Gonto: Exactly, I think it's the same, like, I would never use ads before PMF, at least to get people to come to your product, to your website. The only way that I would use ads, that I would use ads before PMF, is for message testing. So sometimes it's cheap, where if you want to target specific people, you do ads Uh, that are very different one to each other and you see which one people click the most and that way you can potentially test some messaging to then put on your home, but still not great before PMF, I think.
Hank: And that's the other thing I see people wasting money on ads with is bad AB test where they'll have like 2 variants of an ad and they go “we're testing we're experimenting” and they change the “the” to an “a” in the other one. So like yeah, that's a type of type of way. So I like what you said very different messaging, you know. One should be… I don't know… Do you have an example?
Gonto: Like what I'm used to it was something like one is maybe more generic one is more specific, or maybe one you talk about I don't know with retool for example with one thing we were trying was help you build internal apps versus help you build business software just to test which one was better Those are so different that you actually get good results from trying it in some way
Hank: Yeah, and it's always good to test like the latest buzzwords Because when you, when you've got marketing leadership coming in, they're going to be interested in testing like, uh, like for, um, CMS is composable is like a big buzzword right now. And I really haven't seen anybody test. Does that language actually work? Does it matter when you say composable do, and you're targeting developers, do they actually know what it means to know to click on it? Or do they just think that's not for me? And that's something you can test with ads.
But generally, unless you're running really good tests, I would say ads are just giving money, giving your VC money right back to Google.
Gonto: Exactly. And I think like on top of that, there's a couple of things that are interesting to me. Like number one is most developers will not click ads because they have ad blocker or something like that. Even now, a lot of browser, well, not Chrome, because Chrome now disallowed a lot of the ad blockers.
But if you use Arc, you use Brave or some of the others, you'll get them mostly disabled by default. So a lot of these ads, Are not even seen. Having said that, I think there's better ways to do ads, which are more maybe sponsorship instead of ads. Like for example, you could sponsor an open source projects so that's, they show you in the website. That's something that we did with clerk with authJS, that if you go to the auth JS website, you'll see that in the hero, it says something like looking for a hostage solution, try Clerk. Similarly on the sidebar in dogs, it says looking for a hostage solution. And if you click there, it will go to Clerk as well.
So thinking about some of these sponsorship, which are ads, but actually are part of the process that the developer would follow. And it's something that cannot be hidden because it's not like a Google ad or something like that, that is basically blocked by the ad blockers. I think it's good. You could also sponsor, like we talked on the other episode, YouTubers, which I think is a better way of sponsoring or even building what I call content products, which is this idea of building a small product that is sort of an ad for what you build.
For example, at Auth0, we built a website called jwt.io. And the idea of that is that you could debug JSON Web Tokens so that you could see which one worked, uh, better, what information it had inside, and why. And when you use that, which developers used in a lot of cases every week, or in some cases every day, you would actually see that it was built by Auth0, which again, wasn't blocked and something like that.
Regardless, I would agree that doing these things that I'm mentioning before PMF makes no sense, but maybe after PMF, these type of ads make more sense than doing a Google ad or something like that.
Hank: Another good way to spend money early on might I would rather spend the dollars on events, sponsorships, like you mentioned, and there's various types of events. A lot, a lot of marketers will hear events and they think, Oh, I have to have a booth and I have to get a three X pipeline return. But early on you can just go to stuff and just be there and hang out and you get the word out a little bit. Maybe you throw some dinners on the side. You know, there's lots of ways you can spend money.
Another one where people throw a lot of money away I think early on is PR Maybe a deeper topic for another time But I just don't think people care about the TechCrunch article that you're gonna get.
Gonto:: That's for another episode But I think PR um, or what I called old school PR makes zero sense. Not just before PMF, but if they're even after PMF, like nobody reads the news as much anymore, unless they are my dad, who is more than 60 something. Otherwise, I don't think people read it.
One last thing I want to mention is when I think about ads, the only way that I see ads really working for a dev tool is what I call outbound to inbound. What that means is that like, you have developers who are trying out your product, they signed up, they activate it, they are playing with it. And now, um, if they like it, they need to convince their bosses. Who might be a director of engineering, an engineering manager, or something like that. In those cases, I have used ads just to trigger the availability bias, um, on people. So what I do in those cases is because I know the developer liked it based on data that I have, and it's likely that they're going to recommend it to their director of engineering. I actually start showing ads, ads to the director of engineering on LinkedIn, where I can actually target to them. Not because I want them to click the ad. I don't give a shit if they click the ad or not. What I want them is to see the brand. Because if they see it, because of availability bias, when the developer goes to his director of engineering and says, Hey, have you tried X? Do you know of X? They will say, Hmm, that company sounds familiar. So yeah, like, let's check it out. So it's more likely they will check it out if it sounds familiar. And familiarity was proven by Daniel Kahneman. That actually comes a lot from this availability bias.
Hank: Yeah, a lot of people actually knock awareness campaigns, but you do have to remember that, you know, if you're going outbound only one in 20 companies are looking for or are ready for a solution at any given time.
So you need to have warmed them up because if the only thing you do is once a year hit them with an email or an inbound message, uh, or, uh, an inmail message is what it's called, that ask for a meeting and they just renewed their, their vendor three months ago, then they're not going to talk to you and they're not going to remember you. They're probably just going to block you.
Gonto: Overall, what I would say is that we agree with Gary's take that ads suck before PMF and after PMF, if you're targeting developers, in most cases, they suck as well.
Hank: I'll put a quick plug here that if people like us talking a little more loosely. Not about specific news or stuff, give us feedback. We want to know if this type of topic and us meandering around it. Because that's just the tip of the iceberg. Which is your segue, Ganto.
Gonto: I love it. Now that I'm going to have a kid, I'm going to be able to do that jokes as well. Um, some of the second topic is around how AWS actually built Iceberg specifically and a lot of S3 for Netflix. There was a tweet this week where one of the guys that was one of the core developers at S3 was talking about how Netflix was by far the biggest user of S3 and that AWS actually built Iceberg specifically for Netflix.
Iceberg is basically an open table format, which helps to with interacting with files as if they were databases and it helps a lot with like large data sizes and doing analytics workload on top of that, which is something that, of course, for Netflix, very important. They are very good for reads. They are not that good for writes, which again for Netflix, it's great because most people are watching TV shows. And then they are learning about how they are watching, what they are watching and all of that, which I think was, was great in, in this sense. And what I loved about it is that even a company that is as huge as AWS actually decided to have a design partner. What's your take on design partners, Hank?
Hank: There's a lot of good things here. One, there's always debates internally at companies about, Oh, should we really build a feature out or a product out for one customer or one big customer. Is that distracting us from the masses and the bigger total addressable market? And this shows a foresight from Amazon on, well, Hey, Netflix is a, while they're kind of one at one, one of one at the time, there were going to be many.
And now, uh, Amazon has tremendously benefited from that. And then on your question about getting a design partner, Yeah, it's critical because a lot of people will build in the dark and they don't know if their product is actually going to work. A lot of companies will even take feedback from a customer and say, we'll go build that.
And then they come back six months later and it's not, or a year later, and it's not exactly what, you know, they've been asked for. And what's cool about this from what I read is, Netflix was deeply involved. Their, their teams were working closely together on the engineering. Both teams are contributing hours, uh, toward this. And that meant the product was exactly what Netflix wanted. And therefore exactly what, you know, Netflix's eventual peers and competitors wanted. And that's perfect for AWS. Um, I, I've done some kind of, uh, design partner stuff. I think you have more experience with that.
Gonto: I've done a lot of stuff with design partners. To me, what's interesting and my controversial thought on this is that even if you're doing a platform that is going to focus on product led growth, meaning you're going to allow people to just sign up freely, you should still start with design partners.
I see so many companies that because they want to be product led growth, they actually don't start with design partners and they just think they need to open the product to the world and get people to try it out.
But I think that's the worst mistake they can make because they don't have anybody that is really close to them that they trust that they can get feedback from. So even if you're doing PLG, I would say you need to pick five to eight design partners. You work with them, not just to make the product better, but also to make the onboarding better because it's very hard to get people to try your products. But it's even harder than if they tried it and they hated it for them to come back. So it's a lot easier to actually have good design partners to focus on building a really good onboarding experience and then open the floodgates to the world so that they can try it out. Instead of just opening and then wishing that it's a good experience.And if it's not, wishing that people will come back.
So I think design partners are key, not just for this. The other thing that to me is very important is how do you pick design partners? Because if you pick design partners that are not similar to the customers you want to have, you will not build the product that you want.
In a lot of cases, I see companies just picking design partners based on the interest they got from the VCs. Or from the other founders or their network or something like that. And when they did that, they end up having a product that they don't want because these companies that they get the intro, they just say, yes, but they are not really the customers they want to have.
So when you're thinking about the product, you should really think in the beginning about who is your ICP, your ideal customer profile, and then actually pick design partners based off of this. What AWS also did, which, because they are AWS, that I think is right it is if you can find companies that are your target customer that are also companies that other people admire, you can use them for a case study, but also like other companies will want to say like, Hey, I want to use what Netflix uses. So then it's like, okay, Netflix is using S3 and Iceberg or AWS. So a lot of people are going there.
So I think focusing, if you can, on some of these companies that people admire, they don't have to be huge, but they have to be a company that people admire technically or related to your product. I think it's something that's, um, it's really important.
Hank: Yeah. So I think a company that has a good brand and logo that you can build a story with, usually that has engineers or a team that wants a little spotlight, Netflix, perfect example. Uh, a company that represents the customers you want to get right away. I don't think Netflix is a good example of that, iIn this case. I think they're a better example of this third thing, which is a company that represents a future market or a growing market. And AWS is really good there.
And on that, on that first thing, something you mentioned is don't just get whoever your investors intro'd you to. Um, I've actually had this happen to me where, uh, not for dev tools, but a lot of like rev, revenue operations tools try to get me as an advisor. And I always tell them, look, I, I'm always at small companies. You don't want me as a person. Cause I do things. You do things weird at the start of a company. I'm never going to be at the scale that's helpful for you. If you take my advice, I'll be really happy and I'll use your tool, but it, it won't get you a bigger market.
And so that's the other thing to be wary of is don't don't build for people that won't give you a current market or a future market or allow you to do co marketing.
Gonto: 100%. Something to mention on this topic from my side is that it's important to also notice that it's not the same to have a design partner and a contractual commitment.
Like a lot of companies want to push for these contractual commitments, where they will sign with you if you implement X feature or something like that. Um, there was actually a tweet about these from DHH this week that I deeply agree with, which is you never take contractual commitments because otherwise your product team will end up implementing contractual commitments instead of the features that you want to have.
In that case, that's where design partners are different. Design partners are a representation of your ICP and they partner with you to build the right things. For the right market. Contractual commitment is a tool that you use to close big customers. And maybe you couldn't close by adding some of those things to the contract.
And that's the thing you shouldn't do. But if you do, just do it with a very, very big ways and not with the rest. So understanding the difference between contractual commitment and design partners, I think is key as well.
Hank: I think that's it on our two main topics. We had one other to mention, which was, we saw another, what we'd call a boring launch.
Um, It was by Warp. Warp is a really cool company, but they did one of these re brand launches where the entire announcement was just, hey, we've got new branding. And I think that's just a huge missed opportunity. They could have focused a lot more on their actual product, why you'd use it, and then tie in … You can still announce the brand, but I think to put all that effort into just saying “Check out our stuff. It looks cooler,” and then any hope that anyone who checks it out is going to convert or be impressed or something. I don't know. It feels more ego driven to me or like a waste or like maybe a false hope in this being important to the users.
Gonto: I think in this one, it was actually worse than Dino. I love Warp. I love the terminal, but the launch was worse because at least Dino had that fun video in this case, the video sucks. It shows the logo and that's it. And then if you look below, the main thing that I do agree with their launch is that developers, and I think it's everybody, deserve well designed tools.
I believe that design is a feature. So in that sense, sure, you should announce it and show it, but then like on the other side, it makes me think like, Oh, the only thing that you're shipping is a brand redesign instead of new features, which maybe I need, like, why is that? What's the difference? So in this case, I agree.
Like they should either highlight something else, or at least wait to launch their new redesign and their new rebranding with something else that is a bit more meaty and juicy. But I think it was a missed opportunity for sure. Um, even more considering that they actually had like 31 retweets, which is a good amount, not as much as Dino, but it was a good amount.
And I think they, they didn't use it to actually tell people what Warp is. Like, if I look at this launch, it never says that it's a terminal, and I don't even know what's basically what Warp is by reading these four tweets of the announcement.
Hank: Yep. They, if they paired it with a launch, they could have waited till their next big thing and paired it. That's what I've done with rebrands before Um, and that's what i'll do in the future.
Gonto: I love this joke that um, There's always a new rebrand when there's a new head of marketing. Which is half true to be honest like most new heads of market like now this website sucka I need to do my own because I knew they want to do their own websites But I don't know if that's the case here, but it was just funny.
Gonto: Um on this one we just want to do as Hank said and mentioned we'd love to get some feedback on it. Does this format make sense? Do you like the idea that we're loosely a bit more with topics just because it's hard to find juicy things every week unless it's Wordpress like Wordpress things keeps on happening every week, which is so much fun But other than that, it's hard sometimes to find some other, uh topics or drama or stuff like that Every single week.
Hank: Yeah, you definitely gravitate toward the drama and the gossip. Um Yeah, we had a couple of topics we left off this time, which weren't as meaty. And I think there's going to be a lot in the coming weeks with big events happening. So we'll have stuff to talk about.
Gonto: Thank you everybody for listening.
Hank: Bye.
Gonto: So if you have any feedback, feel free to tweet either to Hank or to me on any feedback on this.
Code to Market
A podcast where Hank & Gonto discuss the latest in developer marketing.