So I’ve built a new thing, because I’m always making things. I just can’t stop myself, and this is one of my latest ones, and I think it has legs.
inclib.com – Is a domain I’ve had for a while. Originally I was going to use it for an idea I had, which was a centralized global documentation repo, with the name being an abbreviation for “include library.”
Though, I mostly liked it because it was a short and snappy domain name I could remember. Though whenever I’ve given this domain name to people I always end up spelling it because it apparently sounds like I’m speaking a foreign language.
There was nothing there for a long time except some tech tutorials in a WordPress, and I had stopped updating it for years.
Sometime, about a decade ago, a friend and I were working on an idea (shout outs to Joe D.)
We had a solid name/domain for it and everything. It was called “findthatprice.com,” don’t bother going now. I can assure you there’s nothing there.
We were creating a price tracking tool, which actually was working fairly well, save for the fact it was a constant struggle to keep the data updated.
I was a fledgling programmer holding things together with whatever the computer equivalent of frayed twine and prayers is.
We used APIs at that time. Then, they were much more free and available. Amazon’s product and pricing API didn’t require a federal background check to get access (hyperbole), and most APIs for most pricing data we were using for free and allowed more requests than they do today.
I had some pretty naive design in this tool. For instance, I looked up the pricing info in real time every time someone did a search. And we had modules for about 6 APIs. It was a lot slower than anyone probably would have preferred.
I also quickly discovered that sooner than later we were going to run into the request ceiling on several of these APIs, especially Amazon.
These days, I’d scrape all of the information. It’s out there publicly. It’s a little more inconvenient, and your code isn’t as evergreen, because websites do change all the time. And forget doing it in real time. Just update it several times a day by cron. You can always scale that horizontally too and just keep adding servers so you can scrape more and more a day. Easy…ish…peasyish.
The tool was good enough to be a resume item and get my foot in the door to several interviews though. Amateur and useless as it was.
I think price discovery tools are a solid idea. I know I’ve always been dying for one. And it’s not that they don’t exist, it’s that the good ones don’t seem to stick around and the mediocre ones are well…mediocre.
When I was working on Find That price, I created even a small Android phone app that could scan barcodes then instantly pull them up through the site. I was pretty proud of that.
At the time I was worried about competition so much, sites like pricegrabber.com or in the case of this scanning tool, there was a company called Red Laser, who were eventually bought and absorbed by Ebay and poof, seemed to disappear.
A core feature of Find That Price was alerting. We ran scripts every day that could update the prices and shoot you an email if something you were watching was under your price threshold.
Eventually I gave up on it. I don’t remember why or when. I think it was a gradual thing. But it was a labor of months that I believed in.
And I still do. I think price discovery and comparison isn’t where it needs to be, at least not for the public. And I think Find That Price’s competing apps were getting bought up exactly because some companies didn’t want their prices compared against their competition.
This a long preamble to say: I’m working on this idea again. Or at least something like it. Or a small piece of it, or a module.
I like modules. I like little Lego blocks of things that can fit into each other and work cohesively (until they don’t). What I like even more is shivs, quick and dirty scripts written for the moment that do their job and are never seen again, or…get embedded in the major works of things for years and eventually turn into an ugly code base of shivs, but I’m digressing sharply.
I’m always in need of a good problem to solve, and I found a handful of people complaining that there weren’t good price trackers for graphics cards, and it occurs to me that this is a problem in need of solving.
I like doing this for graphics cards because there are not that many of them. The findthatprice.com approach was to boil the ocean.
A term you may have heard if you’ve been taking things offline while you circle back at the end of the day in corporate culture. But it’s a phrase I’ve found apt for many of the problems businesses try to solve, especially startups no one will ever hear from again.
Many think the approach to create a business is to solve the biggest problem possible. And I tend to agree, because it’s good to solve problems, the only better thing than solving a problem is solving an even bigger one.
But big problems have to be solved with the eating the elephant approach, one bite at a time.
I think you could do more for world hunger by feeding the homeless guy down the street right now than carefully planning how to feed everyone in Africa for weeks, and eventually giving up when you fail or driving yourself into debt never even making progress on the problem.
And so that epiphany is my idea. Let me target one corner of a larger problem, one module. Let’s figure out the pricing of graphics cards.
There are a pretty small number of graphics card to be concerned with. The magnitude is in the hundreds, at least of those that anyone is interested in. (People probably don’t want to track the price of the ATI All-In-Wonder from 1996)
People do need them and care about the price of these things. Including very large companies paying for server level graphics cards for nearly industrial applications. And when you need to buy 5 cards at 2500 dollars, you might be itching to save a buck.
AI developers, both professional and hobbyist need graphics cards.
In fact, I recently had to buy a graphics card that was almost identical to my 10+ year old relic but with 12gb VRAM so I could do AI development comfortably. Also because it couldn’t fit textures for many modern games.
Crypto currency miners need graphics cards, and gamers, who want the best performance they can, need them.
And I wouldn’t have even thought about this idea if I didn’t find more than a few people complaining about not having good access to pricing information.
I’m hoping the idea SEOs well because I plan for there to be tons of texts about product descriptions and pricing and statistics on the cards.
Not to mention some write ups on finer points of graphics card hardware, and comparisons etc…You know, blog type stuff, comparison tools. All that jazz.
I already have an automated pipeline running once a day to scrape prices from retailers and organize the prices as best they can.
I’m working on the website itself. Right now it’s just a SQLite and PHP mock up with some sprinkles of Vue thrown in. But it’s functional.
But I’m thinking about rebuilding the site itself with an API back end and full Vue front end, eventually migrating to MySQL or Postgres for performance reasons (and because I can foresee a SQLite file getting oversized and out of control quickly)
I also built some nifty charts that track the cheapest price per day for brand and vendor. And I’m adding little features like that day by day.
The long therm plan is to future proof it as much as I can as I have a roadmap where I intend to expand into other computer parts, and then cell phones once I have the process, pipeline, performance scaling, and SEO down for this one tiny piece of it.
The plan is written out to about 2 years, with projections on how I expect the product to scale financially, technically and with its user base.
And I believe this is a project where if I don’t see growth, I have several directions I can expand in to try to encourage it.
I can add features, I can add products, I can advertise, get feedback and optimize search.
And above all, I think it’s something that will be useful. Updates to come.