From Failure to 5 Figure MRR -6 Mistakes Learned From a Failed Startup

0:00hi I'm Rob Walling co-founder of microcomp you're about to watch a pretty incredible talk from a pretty incredible founder this talk is titled from failure to five figure mrr the six mistakes we learned from our failed startup it's by Pierre de Wolfe he's the co-founder of scrapingbee and if you've been following their Journey on their blog and on Twitter you know that they bootstrapped and then raised a very small round from Tiny Seed and it dramatically accelerated their growth and I believe the last time they were public about their AR they were crossing 1.5 million ARR and they did that relatively quickly these are experienced smart

0:36entrepreneurs and Pierre shares a lot of what he's learned building his business it's a very good talk and I hope you enjoy it as much as I did but before we dive into that we have the next micro comp Europe that's going to be in Malta and you should head to microconf.com upcoming Dash events if you want to check it out or you can look in the description of this video for a link to the site I'd love to see you at a future input person event and let's dive in to this talk from Pierre de wolf I'm really excited to be here today at microcomp

1:06and I'm going to talk to you a bit about my journey so from failure to five figure mrr the six mistakes we learned from our first thrilled startup so oh sorry yeah so my name is spear I'm French and five years ago with my lifelong friend Kevin we've decided to go the whole bootstrapping in the Acker Road So yeah thank you we started this whole thing by building a few products three in the last five years and two wheel SAS and the first class we've built was a complete failure so we probably did every mistake we could have made but we learn a lot from it

2:03and today what I want to share with you is basically everything I wish I knew before we began so yeah okay so a bit of context here so the first product we've ever made was shop to list it was a simple Chrome extension which allows you to add product from lots of e-commerce and get notified as soon as a price drop so like pocket but for e-commerce you know it didn't work out at all we got a small traction but we were not able to make money at all from it but we notice this one thing interesting like people who use shop to list were e-commerce owners who were actually

2:49adding their competitors product on the tool to receive personality price drop notification email such as oh your competitors product is now 10 less expensive than it was before we and we thought okay there's maybe a product to build here so we quit our job at that time shop to list was a a site project we quit our job and went full time on pricingbot which was a price monitoring SAS we managed to get a bit of traffic but we never really managed to get lots of customers which was the initial goal you know and after a year we were hovering at 450 mrr so we gave up

3:38and this idea to work on scraping me which is a web scraping API we got the ID because we did a lot of web scraping while working on pricing but and with scraping the finally we managed to to have something working and we think a lot of the good choice we've made with scraping we came from all the mistakes we've made with pricing but so let's begin with a failure that that is so much so the first one was to think that a user is the same as a customer so when you build something especially for the first time it's really easy to focus on the wrong

4:22metrics and it's really easy to see every kind of growing metrics as some kind of validation so one good example is like product end so you launch on product and then you get 200 of adverts you get some press coverage you get maybe featured as product of the day and you're like oh maybe I'm on to something and you're gonna spend a lot of time trying to gather feedback from all those new trials and all but the thing is it's not validation at all and until you can find one two or three people who actually use your product and eventually pay for it you haven't validated anything and that

5:10was a big mistake we've made with pricingbot early on we were really happy in the early days because we broadcasted we send the landing page to many many groups user we did a nice predicted launch we had lots of sign ups we gathered lots of feedback and we thought okay we're into something it's going to work but we were actually lost inside a lot of noises because we haven't managed to get the first single customer as soon as possible like actually our product was useless on scraping me we try to do things a bit differently we try to narrowed our audience a bit in the early days we didn't care about

6:03having 500 trials or 100 produced and what or lots of coverage instead we try to focus and to show the product on and much more targeted audience what it meant that we went to small growth hacker forum for example or wet draping Community or developer community and of course we had less and less trials or traffic than we expressing but we managed to get the first few customer much more easily and much more earlier which was something that changed everything for the rest of the journey the second mistake we've made was that believing knowing your industry is the same as knowing your customer so we got the idea of building pricing but

6:59when we noticed that people somehow some e-commerce owners wanted to monitor their competitors prices right so we thought that it was a bulletproof ID that okay they want to do it they will pay to do it and we were completely wrong so the way it works so with Kevin we knew nothing about e-commerce at all before working on racing but so we learned a lot we built this big knowledge map about e-commerce about Amazon FBA Shopify which was Raising at the time drop shipping and all and we learn a lot but we ended up knowing a lot about the industry and not its player and that was a key issue

7:47because I think nothing replaced talking with a real stakeholder real actors of the industry and so we ended up having a lot of assumption about e-commerce that proved to be completely wrong so basically what I'm trying to say here is that we should have had talked way more to user and when we build scraping B we add this in mind from day one so one small thing we did with scraping B to talk to way more users from day one was to actually let them come to us so when you launch your product it can be hard to convince people to talk to you you will reach out to them by email but

8:31most of them won't even read the email so instead what we did was okay you know what you subscribe you created a trail account so if you agreed to talk 15 minutes with us we'll give you 10 000 Pre-K credits to give you an ID the basic free plan was 103k credit so it was kind of a nice incentive while not I'm not seeing our company's banking account and um it allowed us to have hundreds and hundreds not hundreds and hundreds 150 customer conversation very early with people who actually wanted to use our product more of course when you do that you'll have some kind of abuse people will be yeah

9:20yeah yeah yeah okay give me the free credit yeah yeah bye okay bye those are the game feedback is easy to discard when that happens but then you're on the phone or on Zoom what do you ask so basically here is just a very quick recap of the month test by Rob Fitzpatrick but the main idea is don't talk about you when you talk to customers and never ever give them the opportunity to hurt your feelings because if you do they won't and they won't give honest feedback so instead of asking questions such as what do you think about this idea or do you think I'm on to something

10:04ask things such as how do you solve this or what do you think about this problem and actually the most important question here is what else have you tried for pricing but I remember when we ask that question to some e-commerce honors we heard things such as well you know I've never really managed to do price monitoring so I'm really looking forward to use your product or things such as well it's too complicated we don't do it and we thought it was an opportunity here because we thought well we have a solution actually we had a solution to a problem they didn't care about because of course when you start

10:55something or be the company you'll have hundreds of problem and there will be problems you want to solve and other you want to discard and pricingbot was probably solving a problem nice to solve not a really important one with scraping B on the other hand we heard things such as well we've built some in-house tools it's how to maintain or while we're currently spending five thousand dollars on web scraping and we could really get some cheaper alternative and here we are an opportunity to offer a solution to a problem those people were already solving and were already spending a lot of resources to solve next problem was again in the early days

11:51you don't have that many feedback and you try to know what to build next so what feature what your user would like to to get into the product so there is easy mode where you talk to people and they precisely want precisely know what they want but again because we didn't manage to talk to that many people in the early days instead we're scraping me what we decided to do was to let them give us some clue so what we did for example on our documentation page we installed our jar we later switch to Microsoft colority for those who don't know it's a small SAS tool who can give you lots of

12:36insights on our user precisely interact with your page we created a clear table of content on our documentation and for each feature we were considering adding we added a section on the table of content if people clicked on a section they would get redirected to a coming soon paragraph for example we were considering adding a scription feature to our API and we wonder well do users really need it it would have been slow and hard to get that answer directly especially in the early days so we did just that that is a screenshot section on the table of content and then we waited a bit and we just had to see if people

13:24were interested in it by working on the Red Dot so that's a heat map of the documentation so each Red Dot I mean that people clicked a lot on that button that's not a bulletproof method of course because maybe people clicked on screenshot because they simply didn't know what screenshot were or maybe they are they had something completely different in mind when clicking on this but we found that it was a good proxy to kind of gouging interest of feature or Direction your product can take especially again in the early days the first failure the third one was to think that if you need 30 emails to sell

14:09your product problem is your email so this one obviously applies mostly to low touch SAS the one where you're trying to get people to organically go on your landing page create a trial user product and then upgrade to a bait plan but what happened with pricing both was the we thought as technical founder that sells it had to be tedious and it had to be hard and slow and so when we did some consumer or sales call we ended up in month-long month-long exchange with people who would just never ever sign up but they gave us hope every time telling us things like yeah impressive you know what I'll talk

15:05about it with Michael funder and we'll probably sign up in a few weeks and then a week later uh you know what it was Thanksgiving so yeah let's talk again in a few weeks and weeks later actually that time it's a very bad time for us to sign up so maybe we should wait a bit more and actually here those people they simply didn't want to hurt our feelings telling us you know what your product is crap I'm never going to sign up they didn't want to tell that but we didn't know that and so we just waited more and more and thought okay if we improve our emails eventually it's

15:44going to work out it never did and instead we scraping B we thought okay we can't afford to spend that many times selling our product so if we really solve a problem that people are having now and want to be solved quickly they should sign up quickly and we're scraping the winner test that in 10 minutes on the phone or by chat we were able to sell several hundreds per month subscription and it shows that really is a state of things you want to reach a point where you don't have to fight for every sale you need to make Forest failure was to try every single

16:37acquisition Channel out there so again we were technical co-founder we didn't know much about marketing and when you try to search about marketing you'll quickly find almost an infinite way of doing it and so it's really easy to get lost amongst all those acquisition channels out there and what's hard is that at the beginning when you're in the experimental phase it's really hard to know if it's working or not and usually what happened is that at the beginning the acquisition is not working that well but you can get small results and then you start to get good at it after a few months like you start to

17:26master the channel and then finally you get this whole company interest thing and you start to have exponential results well problem with pricing battery is was that we get stuck in the experimental phase and what we thought was bad results were actually completely natural and so instead of staying focused on that channel we just change every time we go to board or every time we thought it didn't work so we trade SEO published a few blog posts and it worked a bit but we thought well it's not working well enough so let's try affiliation instead was not working either at least not as well as we hoped so we tried

18:13called Outreach and social marketing and ready to add and Google ads and basically almost everything and instead with scraping B we try to focus day one and SEO and we only did just that so same as before in the early days we got only maybe a few hundreds views on our articles which converted to just a few trails and to eventually want customers but we thought okay let's do it again and again and again our luck here was that Kevin had some experience with it because before working on pricing but he had a web scraping blog about Java so that helped a lot and we thought okay we should just

18:57play on our strength and focus on SEO all in and even though we won't ever be the best at SEO at least it will always be better to just keep improving on it and just stop switching around on every possible acquisition strategy out there the fifth fail was to believe in several bullet so what happened with pricing bot was we often we often thought that for complex solution there was one simple problem one simple tool one simple growth hack that could fix everything foreign example of conversation I had with Kevin it went like this okay up here I think we have an acquisition problem yeah we do

19:51so you know what there's lots of blogger and YouTubers about e-commerce I think I found it we just have to do affiliation it's going to be very easy we find partner they sell pricing God for us and we're rich and after a few weeks we had two acquisition problems because first we had to still find customers and second we had to find partner and still we didn't even know how to solve pricing but and yet we thought we could find people and teach them how to Celebration but to others so the idea behind this is that behind every if you think there's a silver bullet who

20:34can think fix something complex it's probably because you're trying to avoid the big important question I see that what we did was Breaking Bad last question for pressing but for example where what we should have asked ourselves was instead is my product useful do we get some usage and of course the answer was no like we would get some trial but people wouldn't choose it at all and then do users pay or enough money again they didn't pay at all so of course not enough money anyway and can I get more user can I find a sustainable and scalable way to increase my user base and again here problem was

21:25now and how do we took a stack back and ask ourselves those questions we probably would have wasted way less time trying all those tweaks and acts and Silver Bullet things the sixth one was thinking consistency was enough so you'll hear a lot things such as success is just showing up or consistency will make you win no matter what but the truth was and I know how obvious it's done right now but at that time for us at least it was not that if you're consistent at doing the wrong thing we're just going to do the wrong thing all your life and I think what makes it hard to

22:19get away from this consistency I told her thing is that in the early days it's really hard to make the difference between something that's not working and something that's working just a bit something that's worth fighting for and something you should really just give up and move move on to something else and so of course if you zoom out and think in months of years well results or no results would show and that happened with spacing but a lot so we thought okay things are aren't great but if we keep doing some manual onboarding if we keep talking to users if we keep doing some content and spending some time

23:09we'll reach some success because we're being consistent what's going on one example of it was Amit came to us one day he subscribed to pricingbot Amit was selling adults lots of them and if I'm telling you this he was because he was selling 700 and needed to monitor six competitors and the way pressing but work was you had to manually match your product with your competitors to get the kind of notification that said okay your competitors product is not 10 percent less expensive than yours and we offer to admit to do the manual matching ourselves so for one full new week 10 hours a day with Kevin we ended up

23:52matching at your dolls on our full screen all day long that was not the best week of my life and what happened is that he actually paid six hundred dollars for one month but actually he paid out of pt4s because he was like oh gosh it should have been her hell to do this so okay I'm going to pay but it stopped right here because two weeks later he told us yeah it's not that useful sorry guys any console and here where we should have just stopped take a stack back and ask ourselves what was going on we just continued to do it so we did it with dog

24:36food mattresses light bulbs and all those kind of exciting stuff and it never worked which led me to my last point which was better quit too soon than too late so it's actually very personal here only coming from our Panic wall to our three company but what we noticed was that with pricing but we quit after nine months but we also saw many people getting stuck at doing the wrong thing for years and years and years and we think it's easy to get stuck because there is always an infinite quantity of stuff you can try to make things work there will always be some acquisition channel to try some feature to develop

25:34or I don't know some yeah some stuff to do and instead what we should have known at least was to ask ourselves always those three questions when building pricing bot which were do I know what I need to do to make it work like you do you like honestly do you think you you know what you need to do and we're spacing about to answer from day one at least stage day five let's say was no we didn't know like we were just constantly flowing in a dangerous stream of experiment and we're getting nowhere second question was am I able to do it so for example we spray him but

26:29we didn't know what we needed to do but we knew that some price monitoring tool were actually quite successful but those tools they were targeted at very big e-commerce monitoring thousands of skus and we didn't know we didn't know how to do it because it meant really developing and all different products managing like technique Kelly much more complex than what we were doing and we were not able to do it but more importantly we didn't want to do it because it meant building a company around the product we didn't want to build it probably would have meant raising some money and hiring more engineers and this is something we didn't want to

27:17do so as we asked ourselves those three questions early on we probably would have win and I save a couple of months so if I had to to just keep to listen from this whole pricing bot mess and then scraping me was to never forget that the only true validation is a paid customer everything else is just left product and complement from external user or it just really doesn't matter and to never forget to take a step back and ask yourselves the right questions when you're building your product so if you feel something is wrong there's probably something wrong and dig a Little Deeper do not get

28:09distracted by this new shiny acquisition technique out there so that's it so thank you everyone for your attention thank you rob Alexander for the opportunity Braden for setting these all things up and yeah if you have any question you can reach me Twitter email or right here and it's not working anymore but yeah thank you thank you hello thanks for the talk um it was really interesting because you mentioned a couple of times that your customers didn't find the product useful and I wondered what the what the leap was between having a product which maybe they didn't find useful and actually being really successful now was

28:53I'm sorry what was that well it sounds obviously successful now but when you're giving your talk you were describing that your customers didn't find the product useful so what was the leap between not being useful and being useful now yeah so actually the problem was with the previous company pricing bot so here is the thing user we're not finding it useful at all and with scraping me because we early on really focused on getting that for customer as soon as possible we didn't care about getting traffic or ready to vote or predictant of what we just monitor usage a lot like the first few days after launch I was

29:31just on data dog reading every single API call trying to understand how people would use the product and it's the kind of things like when it happens you know it like when people sign up and spend one hour train your product and then send you emails to understand why is it not working then you run to something because you've built something people will fight to use to to get value out of it hi I'd love your slide of the two graphs of the strategies of working and not working that in the early stage it looks the same yeah uh it's only until later that it works this is a problem that I

30:12have all the time of like how how long do I stick with something to to see if it's working uh so I'm I'm curious I don't think I don't think I quite understood how you tell the difference how long do you stick with something before you realize it's a good strategy or that you just didn't give it enough time I think it's very personal here so for example we spray think about we know we were at the time with no matter what we did we never managed to move the needle even a bit so with pricing but it was actually a complete flat line so I think

30:46if you at least manage to create a slope and if you do some math and if you think you can increase the slope in a few months then you should probably keep going and if you like it also because it's not only about numbers like if you're doing something you ate but which is not also very effective you should probably quit but we can maybe talk about it later if you want like because it really depends if it's an acquisition issue or usage issue or onboarding issue so yeah happy to to talk more about it later you on your dashboard you use this promo 10K API keys for a phone call with you

31:28yeah are you still doing that and is it still worth your time no not at all because after a few months you just ended up getting the same feedback and the same feeds back and you're here always the same thing and because thankfully we managed to get some traction like Kevin was fed up getting maybe something 30 calls in a week which would go nowhere so we stopped doing it at least what we do now is like when people sign up they get to fill a small form and we ask them questions such as how many API calls do you plan to make and if they want to make a lot of APA

32:07call and if their domain name is legit so not Gmail not whatever we give them the opportunity to schedule a call with us hey Pierre hi hi thanks for your talk great thank you um I was wondering about uh for pricing what you said okay uh if we would have built a tool which would monitor uh thousands of skus then it might be useful for those people but we didn't want to do that um but for scraping B with these amounts of customers and these amounts of pages being script I guess your products is also becoming quite complex or at least the infrastructure what was the

32:46difference between those two so actually that's interesting interesting because so is pricing but we are to scrape thousands of website and also for each product page extract the correct information such as price name and it can be easily quite complex so there's easy use case so I don't know some video game on Amazon but then there's uh let's if you want to buy a mattress and then you go on the page and then you have like 20 different sizes and for if you want the price for each sizes you have to click on the button and all so actually scraping me is a much easier product to build than pricing but in

33:27that sense because we just we're a data provider and then our customer extract information from our data and with pricing what we were an information provider so yeah that's why all right thanks for the talk um you can put a really good slide on about how you sort of gauged feedback from users about what to build next with your hop jar dots yeah in the early days in terms of is your sort of success Guru and your number of customers grew did you still keep that tactic going or did it change to something more sophisticated no honestly we keep doing it so now what we often do of course is

34:03like we're thinking about building a feature and then we ask people about it because now we have enough user and for every image email we send we managed to get 20 30 answer but I'm still doing it because what I like is that I can test all the craziest ID here for zero cost and usually Kevin don't even see it so he's not mad about it when I write some really weird stuff on the documentation and sometimes it's a failure like no one click on it and sometimes it works and ajar and a Microsoft Clarity are really well made tool which allows you to like discard for example on the documentation

34:45you can discard data from people who just spent five seconds on the documentation for example so what was useful for example was to only grab the information for people who spend two minutes on the documentation people who really wanted to get in or people who already add an account so yeah we're still doing it hi beer a really cool story thank you um my question is about you about the future so if you look forward the year from now will you keep doing the same strategy uh product wise marketing wise sticking with SEO or you will start adding on top of that something else so it has a lot to do

35:27with how are we going to grow the team so obviously right now not obviously right now it's only Kevin and I with Etienne doing some development and Kevin wouldn't have time to trade something else so either we hire someone to do for example called Outreach or we work with a cold Outreach agency but for now at least in SEO there's so many other opportunity out there I'm just talking in terms of you know keywords and search traffic that we feel we have at least six months one year of SEO we can do right now work on SEO for six months and still uh get great results

36:13and uh yeah that's that answer okay cool response to a year left so basically we have a list of Articles we want to write and um we know that we can write good articles at a rate of four to six a month and that's how we did you know with a very complex math equation we managed to get to two six months okay so your double down on SEO and you're writing a ton of Articles how do you figure out what's the good thing to write about are you basically thinking of what you think will make sense for the customers what's valuable or you go for example into age

37:00reps and look at the keyword support and what people are searching for and have like more data data driven approach so in the early days we didn't even use ahref um in the early days we just learned about this technique called the skyscraper SEO technique so basically the idea is like you write the best possible content usually educational content about your industry not it doesn't have to be about the problem you're solving so for example for us it was about swept scraping so we wrote very detailed with web scraping guide in Python for example and then what happened is that if the content is good you're going to get some

37:44a traffic and then backlinks more backlinks mean are your domain rate and more and a higher domain rate mean that then you can write those specific articles and rank on Google quicker and faster but after a few months so we did that for three or four language and after a few months what we actually I forgot to mention it on the call on the talk but we followed the ahref course blogging for business which used to be 700 but which is not free since last year and thus helped tremendously and from that point we started to to do more serious stuff using href training to know

38:28uh volume and more importantly keywords difficulty because then we noticed okay these articles we let's not write about it because like the first one the the website ranking first on it is like pie charm or something like that so we can we can't even fight so yeah great talk thank you thanks um I have a question about the building public thing which you were doing yeah how has that helped you if anything or is this just for fun and all your customers are coming from SEO or has Twitter helped at all in terms of customer it hasn't helped at all so we currently have five four four hundred

39:19500 customers maybe one came out of Twitter but what it did help with was you know like talking here for example so I know Alexander and Rob got the ID from my Twitter thread and then last time also I tweet about using Microsoft clarity you know and you know and in a week there is going to be a uh use not a use case a case study published on microsoft.com website with a link to scraping me so you know it's a lot of small seed that somehow sometimes flourish but yeah using this to get customer I wouldn't advise it at least if there is not a big

40:04match between your audience and your products so as I built a Twitter tool for example it would have meant make much more sense to building public as an acquisition strategy so it brings a lot of good I really like it it makes me meet lots of very interesting people but it did not help grow scraping bee business at all right that's our last question everyone give PR another round of applause thank you so much here thank you everyone thank you [Music]

40:49[Music]