deploynets.ai

Karma: 32
Recent posts
Recent comments
deploynets.ai · 417d ago

That's a significant improvement, congrats! It looks much better now in my opinion. Some feedback still:
- Keep on working on your messaging. Your first sentence especially ('Slow Code build, compile cycles taking toll on productivity') isn't very clear yet IMO. Know that even big startups constantly seek to improve their messaging so you're not alone in that struggle
- Make the call to action in the first section ('get early access') way bigger and with more contrast (a black background and white text for example). Also try to center it relative to the block of text above.
- Your FAQ link is broken
- Put call to actions in your pricing cards
- Try to have working links, like your terms of service, privacy policy and social media links. You can find templates online for the first two. Not much work and it's always weird to click on a link that does nothing.

As I said, great improvements relative to the first version. Keep working on it, your landing page is very important to have a good funnel start :)

deploynets.ai · 418d ago

I'm not sure why you experience such low conversion rates but here's some feedback on your landing page, I hope it will be able to help.
- The general feel: it doesn't feel professional, mainly because the UI is a bit awkward. I'll break down what I mean by that below.
- The first section is very very empty, only a video and a button. So the only info I have is the thumbnail of the video ('compile code faster') which doesn't really tell me what your product does. As a matter of fact I'm not even sure what kind of website I'm on, it's not clear you're selling software.
- Most people won't watch the video at first, you have to interest them enough in your product first.
- So: first visitors don't watch the video at first (too much friction) and don't really know what your product is ('compile code faster' isn't descriptive enough). So they scroll down to the second section, which is a bit chunk of text in a small font. They don't want to read that. Too much text, too hard to read.
- So they scroll down to the third section (the font is still too small). What you're trying to convey here is clearer (memory efficient, compute efficient, and fast), but I still don't really know what you're actually building. Because I skipped the video and the block of text.
- The video itself is a 2:30 screencast which isn't very enticing either.

Here are some advice to improve your landing page:
- Look up some other software landing pages and emulate them. You don't have to reinvent the wheel for your landing page, focus your innovation in your product. Think of 3-5 SaaS you know, take a look at their landing page and copy them. You can also use website builder services with templates, that'll save you some valuable time (webflow, wix, ...). Some examples of good landing pages: mixpanel.com and segment.io
- Think a lot about your messaging, and try to be very clear and concise about what you're building. On the two landing pages I shared above, as early as the first section I'm very very clear on what they're building. I also know it's a website trying to sell me software. Here you can use Pioneer to your advantage, ask for feedback on wether your messaging is clear enough in your weekly updates.
- Clicking on the 'get early access` button gets me to a 9 questions form. That's way, way too much friction. Collect the minimal information you need, you'll do the survey later. Or better yet: get visitors to use your product without any form if possible. If you're really adding value they'll sign up anyway later on.
- Be clear on the pricing of your product. Is it free? If so, say it (and maybe address your business model, free products can make people suspicious). If not, have a section explaining your pricing.
- Your product is between the devs and the packages they use, a very good place to insert malware. So you have to address privacy and security. Is your software open source? If not, how will you guarantee to your users their security?

I hope this feedback helped. Your product seems to be solving a real problem, but I think can improve a lot on your communication/marketing. That's also a good news: there's lots of low-hanging fruits that once done will yield good rewards.

deploynets.ai · 420d ago

Thanks! Which project do you mean? If it's the one mentioned above I never ended up doing it, it was between this and the one I'm doing right now :)

deploynets.ai · 420d ago

Very cool, I nearly started to work on a very similar idea. Curious to see where you'll end up (and rooting for you). Feel free to sign me up if you have a newsletter or something similar: pierreouannes[at]gmail[dot]com

deploynets.ai · 469d ago

Hey there, my two cents: GPT-3 is very very impressive but I think you'll have trouble making it work for such a constrained environment. This is not something he will have seen during training at all, and I'm not sure the few examples you can give it to make it understand your task would be enough to do what you want to do. If you could fine-tune it with dozens/hundreds of examples then it could work well I think.

However I think that your goals are totally doable with current NLP methods. GPT-3 has been making waves but there's plenty of other good models for different tasks in the field :)

deploynets.ai · 476d ago

Here it is as a jupyter notebook gist: https://gist.github.com/pouannes/1239de4cbb9f16136713039c88156cfc

The list is at the bottom (last run was at the time of this post). I have not excluded past Pioneers though.

deploynets.ai · 478d ago

I have a script that I run when I'm curious about the leaderboard that gives me the top100 of the leaderboard. Basically you go to each region to see the top 30 and aggregate the data to get the global leaderboard.

If folks are interested I could share the (python) code I use to scrap the data and output the global top100

deploynets.ai · 504d ago

Definitely head over to startupschool.org, one of the best source of information on the entrepreneurial world.

This next bit kinda of depends on the time you have. If you have some time, I would advice you to learn to code. Not to build everything yourself, but if you're not technical you'll have to speak and work with a technical person. And when that happens, knowing a bit about how to code will help you evaluate this person and better communicate with her, which is crucial.

Shared by karen@palabra.io · 506d ago · 22 comments
deploynets.ai · 505d ago

Starting with the problem doesn't mean you're not dreaming. You're dreaming about solving the problem, instead of dreaming about your idea.

I should mention that what I said above only applies when you want to build something people want. If you just want to work on your cool idea, great, do that. Maybe it'll turn out that some people really want it, maybe not, that's fine.

Also I'm not saying that starting with an idea never works, or that starting with the problem is The Right Way (TM). Just that if you're looking for good ideas, starting with problems is a really good way to find some.

deploynets.ai · 506d ago

It's the difference between having ideas to solve actual problems you're having/seeing, or to try to come up with ideas that sound good and then look for a problem that could fit that idea.