Pre-selling a SaaS, without knowing if you can technically build it?
Hello my dear Indie Hackers,
Working on a SaaS project, I am obviously wondering about finding several dozen (identified) customers to pre-sell my solution. This, to validate the product and build serenely.
However, I am not sure that I can technically produce this product. To make it short, it is about automating actions on platforms that do not offer an API.
Using Javascript libraries like Puppeteer makes this kind of thing much easier today, but platforms are increasingly protecting themselves against these automations. There are many variables that could affect the functioning of the product.
So far, I've managed to talk to enough potential clients (and also thanks to my experience) to I think, validate the interest of the idea and start creating.
What do you think about it?
Have a nice day! :-)
One approach to this is to keep track of how many people click the 'Buy Now' button without ever taking their payment details. Instead, just show them a message along the lines of 'Thanks for your interest, we'll let you know when the product is ready'.
The upside of this approach.
Of course, there's a few downsides to this approach.
Personally, I've never tried this approach myself, but I've read about others doing it with some success.
Hi!
Very interesting approach, indeed the "trick" could be misunderstood.
After reflection, I think I can find a compromise.
The idea would be to propose a free registration, giving access to a reserved place for the launch, as well as a paid registration, pre-sale type, giving the right to 3 months of subscription for the price of one as well as a 1:1 onboarding.
Thanks for this idea!
Would you refund your customers if your MVP never materialized? Personally, I like how Gumroad does preorders—I pay when you launch—automatically.
Yes, absolutely!
The purpose of these pre-sales is only to give me the opportunity to really validate the idea.
I didn't know that Gumroad allows you to do that, I'm looking into it right now!
I'll look into it right away! Thank you!
Hi,
I think that sounds like an amazing way of validating your saas. Things that seem hard to build or hard to scale are always really good problems to tackle.
What I would do first is doing it manually. If you can do it in a manual way (e.g. going to the website, entering the data somewhere, and sending a JSON file) you can really validate if customers stay and if they pay.
This is of course only feasible for a handful of tasks and with enough capacity. But if it is working like that you're definitely validated enough.
For scraping there are many hosted solutions already where you don't need to take care of the basics like captcha, IP blocks, etc.
Hi!
The automations and scrapping actions will be done on my clients accounts on marketplaces, I can't really do it manually.
The idea of this project comes from my own experience on these marketplaces, which do not allow me to automate repetitive tasks or to graphically visualize my sales / profitability figures.
Yes, I recently discovered Browserless, a service that allows you to "delegate" the infrastructure needed to run a Puppeteer in the cloud. The cost is relatively low and brings a lot of advantages in stability and scale. Even stealth mode is available.
Cheers!
Pre-sell the saas
Pay for a dev to build it for you
???
Profit.
Hello dear Robin,
If I get it right, you'll build custom scrapers for specific actions. Let's say a headless browser will visit IH to check product revenues upon customer's request.
It's not impossible but it'll eventually turn into a cat and mouse game. With many cats and mouses jumping around at the same time. IH will change a variable and you'll start over to figure out what's going on. Meanwhile this action will be down,
how long those customers will happily wait for you to make the fix.
how many scrapers can you fix a day while building new ones.
how many scrapers do you need to call it a success.
how much do people suppose to pay monthly while it's offered dirt cheap on global freelancer sites.
I'm not expecting the answers, I'd ask these questions to myself in case I'd try something like this. Oh wait a second, I may actually do. :)
Overall it doesn't sound like a Software as a Service at least it heavily requires human touch with fixed recurring payments. Which is fine I guess. If it takes 10-20 customers to reach ramen profitability and they are waiting for you, then that's great. I'd like to see it works. So good luck.
- A.
Hi A. :-)
You have understood the purpose of the scrapers.
Indeed, paying someone to create the product is tempting; but what about doing the maintenance afterwards and the improvements? I see too often around me people who become dependent on freelancers or agencies and end up having to pay large sums.
Yes, 20 clients allow me to be Ramen Profitable.
Yes, it's a cat and mouse game with these platforms, I understand that.
The scraping workflows will be common to all clients. I imagined creating an automatic refund system with Stripe, refunding clients pro-rata depending on whether the service was down say 12% of the time during the previous month.
Thanks for your feedback!
What are you working on?
They'll still be dependent on a tool, your SaaS and you eventually. Instead of paying when it's broken, they have to pay on a monthly basis.
I spent the last 3 days on upwork looking for scraping gis. No luck so far. There are several people per day looking to update their existing selenium/scrapy or puppeteer scripts. I can't tell if they are genuine or not but it's there.
Also on fiverr there seems to be a demand but competing there is not an option for me. And in some mysterious ways there is not enough search volume about web scraping in google (checked through ahrefs).
What I'm thinking for myself is there are already tens if not hundreds of SaaS's for automated scraping but as a customer you need to pay for the tool and do the job for yourself starting from around $39. I'm thinking about a freelancer approach for people who needs one time scraping. IDK it's still work in progress.
Anyways if you already have those contacts who are willing to pay, just do it in old fashioned way. Start today. Get paid, build their scrapers, do the maintenance. When you are no longer able to do it manually, build your solution or pay for it and scale. 🚀
edit: I forgot to mention; refund is a good policy but if I rely on that data to run my business. Refund is not going to help me keep on working with someone.
Just my two cents...
We did something like this for internal purposes to automate specific work for the team.
The huge problem was 2FA, security challenges, user preferred language, etc.
API is not only about accessing/manipulating data but also about providing the way to access the platform ;-).
Hello,
2FA, security and stability are the biggest challenges yes.
I don't spend any time thinking about IP rotation, proxy etc, for the moment.
How did you solve these challenges?
cheers!
We used the same origin (a computer available in the office) so that the only thing we need to solve was to detect 2FA/security challenge and ask the related person to confirm it and, of course, mark the browser as a trusted one.
Therefore, we used normal Chrome controlled by Selenium, I think. As soon as there were any issues, it was just about turning the monitor on, solving it, and resuming the script. Btw, the computer was used also for other things (mostly running Docker containers), so it was kind of the simplest solution to reuse it for this task too.
Also, since we are talking about 15 people and occasional actions, we were not concerned about traffic and don't use any kind of proxy.
However, it was a real deal to be prepared for all the different edge cases.
Btw, for a completely different job, we scrape some websites and https://stormproxies.com worked pretty well for us.