So now it's like, oh, okay. Now we have to rethink all these different attributes and the meta attributes of those attributes.

Oh, okay. Yeah.

And so what was the solution? Yeah. So the solution that we came up with was this. So all these ETFs and they're these like specialty like browsable subcategories that were made were primarily made for people who were using the automated roboadvisor portfolios, basically. So basically we said, okay, we're making a new stock investing platform. And yes, you can get into ETFs. But if we're really trying to emphasize like far different chain factors, really these like stock collections connected to these stocks, do we really want to try and create this like super in depth like categorization system that tries to subdivide all the types of ETFs and all the different types of stocks and try and make this like super comprehensive filtering process, if the answer to that was definitely not, it just became a matter of like, what is going to be our different factor and where do we want to play? And so the solution was, you know what, we're going to have this like filtered down search space within just the automated portfolios where if you want to like go and edit your automated portfolio, you can. So you can hit like this like edit button, it goes to this like sub search space where you can still find those categories, so you'll get rid of them, they're still there. We can only find them there. And so you have this sub search space for this one group of people and this prevents people from this like top level search space from getting distracted by all these categories that they probably don't care about. So we just made a trade off is basically what came down to is like prioritize for these delegators and this wealth front, automated portfolio space. And then for these more hands on people, keep it focused on the stocks and the collections and that categorization scheme. So it's like, categorization schemes and different spaces. If you want to use that information architecture term. How many collections do you have? How many stock collections are there? What how many instances? I don't know how many. Is it triple digit? No, definitely not triple digit. I think when I left it was probably in the 50 plus. But yeah, we built up systems where it's easy to like, if you want to make some new ones, you have all the processes and place to figure out how to do that. So I create like voice and tone guides and some pretty in depth like content prompts. So people are like struggling to like write a description. I would say like, well, if it if this collection contains this type of like idea or type, then you should think about how to front load the description this way. And here's a visual way to do that. So yeah, we just created a bunch of processes in place so people can keep building on those collections over time and then storing them in these pretty, pretty big spreadsheets. Is there a way for the users right now to sort or filter the collections? There's not. Okay. So this is the one problem that I really had was like, how do you how do we do that? Should we do

that? Because there's actually

no collection level metadata.