Why is all text lowercase on your website? I wouldn't want this in my product :)
Oh! That might tie into my other comment about how this would work even better if users were individually tracked!
Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
11–20 of 80 posts
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#12Why is all text lowercase on your website? I wouldn't want this in my product :)
Another reason (branding) why copy would look different for different products :)
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#13Why is all text lowercase on your website? I wouldn't want this in my product :)
Oh! That might tie into my other comment about how this would work even better if users were individually tracked!
So it would detect you for example as some sort of millennial, and use the punctuation that was passed down to you as correct and a signal of competency
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#14Cool! What if there isn't one optimal copy though. Would this ever be tailored to an individual user (or group that a user is in) based off of information you know about the user? So some sort of tracking knows that your opti-cop worked well on me for some other site you service, so then it tries a similar style for another site (who uses your service)?
Yes this can be tailored for individual users (or groups). To go more in-depth, we'll be using contextual bandits + more novel ways to optimize the best performing copy for user segments, that works on top of the anonymized knowledge about the user.
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#15Neat product. Small thing on the demo: perhaps using Stripe in the demo isn't the most effective choice. Stripe surely has already A/B tested the heck out of their landing page; "Financial Infrastructure for the Internet" is (IMO) an incredibly strong tag line for the hero text. The alternatives generated by your tool pale in comparison. Perhaps it would be more effective to put a lower-quality landing page in as you…
Neither Stripe nor Intuit are converting users from their hero copy at this point. "Financial Infrastructure for the Internet" sounds like a diversification pitch to investors than anything else to me. A good example would be a company that isn't an unavoidable juggernaut in its space and also isn't great at marketing. There are many open-source projects that would work because they don't have a big marketing team, b…
You'd be surprised how few companies actually do any systematic copy optimization - it remains ad hoc even at the big players, which is one of the reasons we started this startup.
In speaking to SMBs and large companies, our insight was that the problem of copy optimization resonates more with larger companies, as smaller companies are more focused on survival/basic marketing techniques like opening up new channels. Larger companies have already exhausted those levers, and are ready for more sophisticated optimizations.
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#16Earlier quoted context omitted.
Oh! That might tie into my other comment about how this would work even better if users were individually tracked!
So it would detect you for example as some sort of millennial, and use the punctuation that was passed down to you as correct and a signal of competency
Yes, we think of this as a decision tree, where initially, copies may look different by say, user demographics and location. As we learn more about what's working well for different dimensions of users (eg: topical interests, traffic source, platform type), the decision tree grows, and every single element in the anatomy of copies is optimized based on past learnings. In an ideal world, every user truly sees a unique copy tailored to them.
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#17I’ve spent lots of time dealing with the pain of this in past growth roles so I’m excited to see this! How are you handling messaging consistency for specific users? Ie. A user is shown an experimental string of copy with one value prop and you want it to show up multi-channel for them. Do you have a way to associate the experiments on a per-user level?
Every copy has a globally unique identifier and multiple copies can be changed in a single experiment, agnostic of channels. That way we only need to setup experiments on user level and as long as different channels are referring to the same experiment, consistent messaging will work out of the box for each copy across channels.
example: website_landing_page_title and email_subject_line can be part of the same experiment for a multi-channel experiment.
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#18Earlier quoted context omitted.
Neither Stripe nor Intuit are converting users from their hero copy at this point. "Financial Infrastructure for the Internet" sounds like a diversification pitch to investors than anything else to me. A good example would be a company that isn't an unavoidable juggernaut in its space and also isn't great at marketing. There are many open-source projects that would work because they don't have a big marketing team, b…
You'd be surprised how few companies actually do any systematic copy optimization - it remains ad hoc even at the big players, which is one of the reasons we started this startup. In speaking to SMBs and large companies, our insight was that the problem of copy optimization resonates more with larger companies, as smaller companies are more focused on survival/basic marketing techniques like opening up new channels.…
> You'd be surprised how few companies actually do any systematic copy optimization - it remains ad hoc even at the big players, which is one of the reasons we started this startup... In speaking to SMBs and large companies...
Did you talk to Stripe? Do they do systemic copy optimization on their landing page, or not?
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#19I'm no lawyer, but the statement "to productize the tooling we started building at Twitter" is troublesome to me. Do you have some protection from Twitter/X making a IP claim against this? I.e., can you prove you scrapped "the tooling we started building at Twitter" and began fresh?
The idea transpired from some of the early tools at Twitter. The actual product has nothing in it that would be considered Twitter's IP afaik. A couple of other companies we spoke to use similar techniques to optimize copy testing. Good call out on the language though - we could have worded it better!
"Good call out on the language though - we could have worded it better!"
A little ironic that this company is all about wording it better :).
Re: Launch HN: Just words (YC W24) – Optimize your product's copy for user growth
#20Meta, TikTok and Google have all the data. Not even a sampling. They can exactly answer what is the most effective copy for possibly anything that anyone wants to do. No A/B test required.
So why don't they offer that?
I know it's not because the data is proprietary or private, because basically all the information you need is visible on Facebook Ad Library, more than enough to answer most questions about authoring copy by sheer mimicry.
You emphasize the UX here a lot. I don't know. I think Meta's UX is fine, end of story.
This isn't meant to be a takedown. It just seems intellectually dishonest. Anyone who has operated these optimization systems at scale knows that the generalized stuff you are saying isn't true. You're emphasizing a story about engineers versus product managers that is sort of fictional, like the right answer is the one that most companies are already taking which is to not A/B test at all, because it doesn't matter, and when you do see results they are kind of fictional.
And anyway, it belies the biggest issue of all, and this is actually symptomatic of Twitter and why things were so dysfunctional there for so long, long before they were taking private: you are saying the very Twitter esque theory that "Every idea has been had, and it's just a matter of testing each one, one by one, and picking the one that performs best." That was the core of their product and engineering theory, and it couldn't be more wrong. I mean if you don't have any experience or knowledge or opinions, why the hell are you working there?
> However, coming up with the copy, testing it, and optimizing different variants across users would take us forever—sometimes months.
The right answer is right in front of you! Don't do it! Don't optimize variants, it's a colossal waste of time!