Episode 3
· 20:16
Hey, everybody. This is Jack from Bitreach, and you're listening to Scaling Dev Tools, the show that investigates how dev tools go from 0 to 1. Few things in life are as hard as raising money and building a dev tool startup. But one of the things that might be harder is not raising money and building a dev tool startup. But that's exactly what today's guest, Michael Christofidis, is doing with PG Mustard.
Jack Bridger:Michael, it's great to have you here today.
Michael Christofides:Thank you so much, Jack. Thanks for having me. It's good to see you.
Jack Bridger:Good to see you too. Michael, could you tell us a bit about where PG Mustard is right now and what your focuses are at the moment?
Michael Christofides:Yeah. Absolutely. It started it was just 2 of us. We both spent periods of time, full time on it. Dave's currently not, and I'm currently doing about 3 or 4 days a week on it.
Michael Christofides:We've been going a few years now, 3 years, a month ago. I recently shared that we've gone past a 100 customers, just passed about 12,000 query plans submitted through the product. He helps people view post request query plans and helps them speed those up. So, yeah, a few little milestones there. It feel kinda feel like small numbers in the grand scheme of things, especially when you look around at other dev tools, but, you know, super proud, really happy with it.
Jack Bridger:That's amazing, Michael. Could you tell us a bit about what you're currently doing in terms of growth and what's keeping you up at night with PG Master at the moment?
Michael Christofides:A good question. Growth wise, it's super tempting to keep I've got 15, 20 features that I'd love to add to the product, and each one of them, I'm convinced will double the size of the business. Deep down, I really know now that that's not true. We're really marketing and distribution limited at the moment. Right now, I'm focused on a couple of things, mostly search engine stuff, so helping people so trying to work out what people are searching for when they have this kind of problem.
Michael Christofides:Get a good blog post in front of them or some helpful documentation. It's a lot of work behind the scenes. It doesn't feel like it improves the product much, but it does help customers who have that exact problem. Then you have to work a reasonable way of mentioning that you also make a product that helps with this kind of thing and hope that a certain small percentage of those, give you a shot right there and then or remember you for later. A lot of content creation, editing, improvements, and a little bit of trying to work out how on earth, Twitter works and how on earth to be authentic on there and share things that are helpful to people or somehow useful or funny or something.
Michael Christofides:I don't know. Those seem to be out of, kind of, the channels that have actually worked a little bit for us in the past, so I'm trying to double down on those. That seems to be the good advice out there.
Jack Bridger:One of the things that I've noticed when I go to PG Mustard, your website, is that you're very ethical. There's no small print. Everything is as kind as it could be. Is that something you've consciously gone after?
Michael Christofides:It feels like a tricky one. Right? Because deep down, there's some people that grow quite quickly by doing things that they're not necessarily black hat, but it's in the grey areas a little bit. I guess even if you did convince me that those were the best ways of growing the business, I think I'd still have quite I like sleep. I don't like, being kept up at night, and I want to build this for the long term.
Michael Christofides:So I think that's probably the biggest reason behind it is that I'd love to run this business for 20, 30, 40 years, hopefully, you know, forever and build a really small team around it that I love working with. I don't think that's what most people want. I think a lot of it comes from that and I'd hate to have a reputation that wasn't that. I think most of it comes from that and then I've convinced myself somewhere deep down that if you're straightforward with people, they really respect that. And I love brands that feel like they'll tell me the worst thing about it upfront, hopefully, or they're a little bit more open.
Michael Christofides:Privacy policy is actually readable in a few minutes, not 30 pages of something I hope I don't read. The pricing straightforward. There's no hidden fees. I can cancel myself without having to phone them. All those things that are just that little bit shady.
Michael Christofides:I personally try and stay away from those products, and I feel really uneasy selling one myself.
Jack Bridger:Have there been any tangible benefits? Let's say some founders are considering this. Has there been any tangible benefits that could persuade them to become whiter than white?
Michael Christofides:That's a great question. I actually don't know. I don't think I've seen anything explicit. I wonder if we've had a lack of problems as a result. We haven't had complaints or chargebacks or I don't think I've given a refund yet.
Michael Christofides:We've offered we offer 60 day refunds. I don't think I've given one yet. So I suspect maybe that's where some of the benefits come in. We get very little custom support. I'd love to get more, to be honest.
Michael Christofides:I think in the early days of a business, support can be such a good source of ideas and getting a real pulse of your business and pulse of your customers. But we get very little of that. I've never considered whether those two things were related, but possibly.
Jack Bridger:You've mentioned previously that a lot of your customers come from word-of-mouth as well.
Michael Christofides:I suspect that is a huge a huge part of it. Maybe though I'd attributed more of that to the product working and the product doing what it said it would do rather than the business being more straightforward. Maybe those are more intertwined than I realized. The product doing what it says it will do and not overpromising is the marketing copy. It is the first few experiences they had.
Michael Christofides:It is actually maybe seeing that we've mentioned how long the privacy policy takes to read or something as stupid as that maybe tips them from using the product, thinking it's alright, and then never mentioning it to anybody to, oh, these people actually seem like maybe a little bit like me, maybe actually a bit tired of all of the marketing nonsense or the trying to trying to pull wool over your eyes' and that tips them into maybe tweeting about it or maybe telling somebody or maybe going above and beyond to share it. Yeah. Really good point on the word-of-mouth. And I think also we're in a Postgres community or that's at least where I thought we were. So you can either have tried to get closer to and I think word-of-mouth matters there too.
Michael Christofides:So you might do a talk for 1 conference and if you're really straightforward with them, tell them exactly what you do know, what you don't know. If you if you do a talk and you you don't try and oversell how much you know and what especially, like, when people ask questions at the end, I think that's the kind of thing where people take note and they recommend you for other things. You're more likely to get your next talk accepted or you're more likely to, behind the scenes, be recommended for something else. Yeah. I suspect, just spot on there, that it's more in the word-of-mouth realm that it's benefited us than anything else.
Jack Bridger:Michael, you have previously worked as a head of customer success at a big what could be described as a dev tools start up, GoCardless, developer focused at least. Do you think this is where some of your approach came from?
Michael Christofides:It's super interesting questions. GoCardless often gets described as a Fintech. GoCardless were very Stripe inspired in terms of paying a a lot of attention to their developer documentation, really investing in the developer experience. There was that side of it. The other side that I really got from GoCardless was how much they loved Postgres.
Michael Christofides:I probably learned more of my kind of user experience, usability, kind of raw skills at a company called Redgate that I worked at, which was very much dev tools. That was even database tools for SQL Server and I ended up running the very small team. So at GoCardless, I think I really saw a kind of different level of execution. They were exceptional team and they really focused on the developer experience, but I don't think the developer was the buyer and that's quite different. I do think they're a little bit more on the Fin side of that Fintech equation, but they're a very, very accomplished tech org within them as well.
Michael Christofides:But I probably did learn a bit more on the customer side there. So customer success is all about talking to the customers, understanding how they use the product, what they like about it, what they dislike about it, why they might leave in a few years' time, which bits of our roadmap they're most excited about, all of those kinds of things. Now I heard from a friend that well, they asked me recently whether our customers were aggressively self-service, and I thought that was a a perfect phrase for it. At Redgate, it was not exactly aggressively self but it was much more self-service. At OCADAS, it was less self-service.
Michael Christofides:There was a bit more hand holding along the way, a little bit more willingness to talk to you as a provider. I guess you're more important to their stack if you do their payments. With the dev tool, you can really help with productivity. Maybe you can really help with education. But if they had to do without you, they could.
Michael Christofides:So you're just that little bit less important to their stack and they're an extremely resourceful bunch, so they could work it out, you know. So maybe that factors into it. But all the things you learn from talking to people, you start to be able to spot patterns of what people say and what they do based, so you start to be able to look at the data and start to be able to think, oh, maybe this is how they felt or maybe this is what they did. And I do enjoy doing the customer calls still. I think probably I learned some of those skills at GoCardless.
Michael Christofides:But the place I picked up, probably the more useful ones were sitting in with professional UX researchers at Redgate, doing those, watching somebody use your product for the first time and trying your hardest to not say anything and just asking them what they think in every stage and taking the blame for every single problem they hit, that kind of thing.
Jack Bridger:So, Michael, stepping back for PG Mustard, what has and hasn't worked in terms of growth?
Michael Christofides:I am gonna start with the positives. And what I think has worked, and it's always pretty tricky to tell, what I think has been working has been content marketing. So we have some blog posts that we've written mostly for the product. So it helps people with performance advice, and we help within the product, but we also link to read more if they, want some extra information. So those blog posts that we wrote for those started to rank for certain search terms exactly when people need the product, which is helpful.
Michael Christofides:On the content side, I also spent quite a few months a couple of years ago creating some explain so Postgres explain documentation because we thought it was a little bit light because I needed it. Right? So I'd find myself googling for a specific term that came up, and there was you know, I'd have to trawl through Stack Overflow answers and hope that there was a good one or I'd end up in the Postgres source code, to be honest. And it's I don't understand c, but it's really well commented. But it felt like if I was having to do that and I had quite, I was quite used to it, people other people that have come across Explain for the first time or maybe they come across it once every few months, it's pretty intimidating.
Michael Christofides:So I spent a while putting together documentation for that. That's starting to rank over time. That kind of long term search play seems really sensible, especially because when developers have a problem, they go to Google. It makes sense. Or they go to a colleague or they go to Google.
Michael Christofides:So we need to kind of tackle that from both directions. 1st step, Google. 2nd step, colleague. So on the colleague side, it's much more about word-of-mouth, and I have no idea how you well, I guess we talked about it briefly already, but the strategy there is tough. Right?
Michael Christofides:It's try and be well known. And what does that mean? We don't have a tonne of money, so we can't sponsor everything left, right, and center. It's also pretty tricky to target exactly who we who we want. It's not just it's not everybody who uses Postgres.
Michael Christofides:It's only people with a certain amount of data with a certain problem set. So not everybody has performance problems. Not everyone has performance problems regularly and getting in front of them at the right time is tricky. I think things like people tweeting about us has been exceptionally useful. We noticed in the kind of days after somebody with a few 1000 followers tweets about us, those sign ups seem to convert at a much higher rate than others.
Michael Christofides:So I think that probably is where more of our customers have come than not is the the word-of-mouth side. But how on earth do you encourage that? I do not know. Or at least in an authentic way. I think it's very easy to incentivize people to tell each other about you via affiliates or shadier tactics, I guess.
Michael Christofides:Most of us now smell that coming a mile off and when somebody's being authentic or when they've got some incentive behind them. Doing that in an authentic way that actually works long term, I don't know the answer to other than make a good product, treat people well, and hopefully, good things come from it. On the side of things that haven't worked so, I I guess, we have tried some paid things. I tried Google Ads for a while. I thought it being a low competitive market, you know, most of our direct competitors are completely free.
Michael Christofides:So they won't be spending any money on Google Ads. So it felt like a good opportunity, but Google had, like, minimum prices and it I think they set out minimum price at about a pound per click or something, which seemed bizarre given we didn't have any competition on those terms. So I don't know how it works. We also don't like tracking people for some of the reasons we've discussed before. So measuring if it actually worked would have been super tricky.
Michael Christofides:Super low volume as well. So just didn't feel worth pursuing. It didn't feel like that was where we were gonna crack the market. And, also, how many devs don't have ad blockers? It just doesn't seem like didn't seem like it would make sense as a business.
Michael Christofides:So then tried some other little things that might work in the future but don't seem to be working now, which is, like, sponsoring an event, sponsored a newsletter, some of those things. And I think probably the mistake I made there was going after postgres related things, which might sound sensible, but I think in hindsight, people that like the product best tend to self identify as full stack developers, and their community tends to be the rails community or the Django community or something else. And maybe they subscribe to a Postgres newsletter as well. Maybe the ones are a bit more into Postgres, but maybe not. And they definitely don't identify in that crowd or they don't go to Postgres conferences.
Michael Christofides:So, yeah, maybe that'll be in our future, but so far, I've not had any success in that or no obvious success in that area.
Jack Bridger:Related to what has and hasn't worked, are there any kind of general lessons that you would give to maybe yourself starting again or another dev tools founder?
Michael Christofides:The thing I think we've messed up the most and for the longest was pricing, and I think it was based on trying not to make the mistakes everybody says you're gonna make. We've been we've both been around software, like, working at software companies for more than 10 years or close to 10 years by the time we started PG. So we thought we knew our staff. I'd been a product manager. I'd done pricing projects before.
Michael Christofides:We knew that we knew the advice, you know, charge more and charge for value. And looking back, I think I've worked out why that ended up being bad advice for us or maybe in general. I'm not sure. So to tell the the story briefly, we started with 2 prices, just monthly and annual. Nothing revolutionary.
Michael Christofides:We had a 7 day free trial, pretty normal. I think it'd be hard in dev tools to not have a free trial. People want to try it, especially we've talked about aggressively self-service. And our prices were €50 a month, I think, €500 per year. And I think we either limited you to 10 users or it wasn't limited.
Michael Christofides:I can't remember which. But the thinking was that most teams have one person on a team that does this stuff, but we thought we could help them spread that across the team. If your team gave all this work to 1 person versus if they gave it to 2, there's not that much extra value in there's definitely not double as much value in that second person and there's definitely not triple as much value for that third person. So we thought, you know, really, we didn't want to try and be clever, you know, But in hindsight but luckily, somebody did subscribe to that annual plan quite quickly. So we got some false confidence, but quite quickly realized that there was something wrong.
Michael Christofides:It took a few months to get a second customer, but, yeah, we were willing to try stuff. Anyway, we got feedback that the people wanted to try it, but they didn't need 10 users. They only had 2. Could they have it for a 5th of the price? Or they didn't need it that much or that often, so could they have some cheaper version?
Michael Christofides:Anyway, we thought we got clever and we realized that really the value metric was how many uses people did. So true value. Right? Then we tracked so maybe for the next few months, we we redid the back end, started counting how many plans people were submitting, moved the trial to a per use limit rather than a time based limit. Loads of things that made loads of sense.
Michael Christofides:We could charge the people more that used it more. Got another couple of customers. But after speaking to a few more, we realized people had no idea. They had no idea how many how often they were going to need this. They overestimated in general.
Michael Christofides:People that came in thought they would use it a lot more than people that were already using it. So I felt like they were massively overestimating how much they would use it. And I think they thought they would get less value from each user. Now this is like a melting pot of disaster for us in terms of people thinking they were going to need to use it a lot. They're not going to get much value each time.
Michael Christofides:So the price was going to be high and they weren't going to get that much value. So that that was terrible for us. And understandably, we didn't get too many customers. I also spoke to at least, what, maybe 2 in the end, that were choosing between plans for about 4 months. So they just didn't make a decision because they didn't know which one to go with.
Michael Christofides:So instead of getting a customer on one of them, we ended up not getting one because they just couldn't work out which one they needed. So we quite quickly changed that and went to quite a cheap price and per user base. And we had a few months of really good success off that. So January, February, March of 2020 were quite good. And it was like, yes.
Michael Christofides:We've cracked it. We can now grow. People like the price. It works for us. So we scrapped monthly.
Michael Christofides:We went annual only. It's, again, not advised by anyone pretty much. But our tool suited it. It wasn't useful every month necessarily for everybody. So having a cheap annual price meant they didn't have to cancel and resubscribe every few months and it meant we could have much more predictable revenue.
Michael Christofides:All of the kind of prevailing wisdom was charge more, charge for value, charge monthly, all those things. And for us, that turned out to be awful. And by trying to think things through from first principles, I think we messed it up more than if we just started cheap. If we'd started super cheap per user, kept it super simple, I think we'd probably be quite a lot further along in terms of revenue. So, yeah, I think it's a tricky one and there are certain standards out there.
Michael Christofides:And the more I've thought about it, the more I think that you only get to pick a few of those that aren't the standard. So every time your pricing is not similar to someone else's or your business model is not similar to the norm, people it it puts a doubt in people's mind and the quicker they can explain that as, oh, that makes sense, the better. So with pricing, charging per user, charging by a metric that they can easily know their number for makes sense. Even if you could kind of rationalize a better value metric, you have to then explain that to them. And I think starting cheap is a good problem to have.
Michael Christofides:Right? Let's say you leave money on the table for your first 100 customers, but your business is only viable once you get to a 1000 or 2,000. That's not a problem. You've already got a load of people telling people how great you are. That's a good problem to have.
Michael Christofides:And there's a possibility that all the advice is charge more solely because the most successful people in the world started off by charging too little. Like, it's possible that that's the anyway, I'm not I'm not confident in that theory, but there's a chance it's survivorship bias. And there's the reason people are out there saying charge more is because charging less at the beginning is a really, really good strategy. I'm not super confident in that. But for us, I think charging less at the beginning would have given us a nice little boost and charging by a metric that people can quickly know their number for and not have to think which plans work for us.
Jack Bridger:That's like a pricing MBA right there. Michael, it's been amazing hearing about your journey with PG Mustard and what you're up to. How can people learn a bit more about you and about PG Mustard?
Michael Christofides:Thank you, Jack. My name's pretty tricky to spell, but we're trying to do more on Twitter. That's probably the best place to start. So at pgmustard is is as simple as that, and you can find me from there quite you can see quite quickly as
Jack Bridger:well. Brilliant. Thanks everyone for listening to Scaling Dev Tools. We'll be back very soon.
Listen to Scaling DevTools using one of many popular podcasting apps or directories.
