Postgres RDS has version 0.5 of the pgvector extension installed by default.
Adding vector support to our app was as easy as enabling the extension and created a table with vector columns. No additional database required and trivial to do mixed queries. Maybe Pinecone and friends have better scalability, but if you need basic vector support, you can do it easily on RDS.
Fair enough. I researched it earlier in the year and it wasn't available.
I stand by my point: we often hear that adding a new tool to your stack is introducing unnecessary complexity, when more often not using the right tool is the path of most complexity.
One of my friends does a productized (fixed price) roadmapping session (1) as part of a larger "build your MPV" service. My understanding is that it works out really well for him and his clients. They get to try him out with a fix priced engagement. The output leads naturally to the next step in a larger engagement. If for whatever reason it doesn't workout the client has useful documentation they can take to someone else.
There are lot of similar type formulas. For larger deals you'd have some form of vender (seller) financing which is what the above really is. It's a type of loan from the seller. For smaller deals that are reasonably competitive you're going to probably have to do an all cash offer.
If you think you're in a good negotiating position or if you've got concerns about the business then a vender take back or an earn out or both should definitely be on the table. It never hurts to ask ;).
Are you asking about valuations? FE has a post on how valuations work. It's pretty simple. Basically a multiple of how much cash you can take out of the business. If that number is small and not growing then the multiple is small. If the number is big and/or growing fast then the multiple will be larger. The key is it's about how much cash the business makes now vs. traditional tech VC land is about... how big this might be some day maybe.
If you're interested in buying a business the people who know how all this stuff works are the private equity people (PE). These guys do this stuff all day long (obviously at much larger scales. If you search you can find a fair amount of info on PE. On a smaller scale there is good info from people researching search funds. Search funds are basically small scale PE. Here's a good starting point: http://www.gsb.stanford.edu/faculty-research/centers-initiat....
From the seller's perspective you've got to have a buyer and buyers at this price point buy on a multiple of SDE vs. development cost. If dev cost is greater than what buyers will pay then as the seller... you're probably going to have to eat it and take less.
Importantly as the buyer a big part of what you're buying is customers that you can talk to and work with vs. just the code. If we thought Codetree would have taken twice as long to build but had zero customers we wouldn't have bought it at half the price. Having paying customers proves that you've bought something that other people value. Buying code just gets you code and you may find out that no one will ever pay to use it.
Having your financing together is going to give you a leg up in a competitive deal for sure. Of course not all deals are competitive so it might not matter. In some cases your bank will give you a letter saying they are willing to lend you $X under reasonable circumstances which might help you get in the door with with sellers by convincing them you can close even if you don't actually have the cash upfront.
See my comment above on the alternative approach. The short answer is to approach the problem the other way. Build a list of businesses you'd buy and start emailing.
kareemm has already answered the questions about financing but I actually think your second question about "where to find these deals" is really the more interesting and important question. The obvious answer is places like FE where we found Codetree but that was a total fluke and it's not the right way to approach it.
The truth of the matter is that many business are for sale if you ask. The right approach is to think about the type of business you want to buy, build a list of all of the ones that fit and to cold email them and see if they'd be interested in selling. Basically the same approach you'd take if you were trying to sell B2B software i.e. Predictable Revenue style.
I can already hear people saying, "No way people don't do that" or "That would never work" but it does and people do. In fact we know a number of people that bought their businesses using that approach and it is exactly what we were about to start doing before Codetree fell in our laps. It's also the way that a lot of PE, VC and search fund deals get done.
I'd bet cold-E-mailing could actually work, but how do you find comps to guesstimate the cost of the business if it's not for sale and not much is known about it. I could imagine E-mailing CLOSELY_HELD_SMALL_PROJECT I think I could run better and getting back, "Sure, I'll sell. We generate $5M in FCF and I'd take 3X that!". Me: Uhhhhh, aww shit.
Remember that an offer to buy is also a sales pitch, after a fashion. Unsolicited offers start to come in almost as soon as you feel like you have room to breathe.
Do your due diligence with financial statements. Any business will file taxes and any corporation will additionally have income statements, balance sheets, etc.
To clarify, I'm asking what are some clever ways to ballpark what a company's financials might be, to pre-screen out ones out of my league before I even contact them. Most companies don't publish anything about their finances at all.
Unfortunately, I don't think you're going to find a good way to ballpark their financials without them supplying them. It's common in the cold email to say something like, "Our sweet spot is companies with an EBIT of $X-$Y". So the person you're contacting knows if they are too big or too small.
For deals of this size you can also look for smoke signals e.g. you're looking to buy something for 200k-300k in EBIT and you go to their about page and find 50 employees you're probably barking up the wrong tree.
It's also worth pointing out that you're looking for one person to say yes. If you hit on folks that aren't a fit and never email you back or laugh you away so what. That's the nature of the game. You just need one.
No different than doing B2B sales and emailing a prospect that is 10x bigger than you expected and only buys IBM and Oracle. A failure to penetrate a single prospect just means you move on to the next one.
Exactly this. I want to present my business as being available for sale. However, it is difficult to start this dialog on HN because people seem to find it spamming to bring the subject up. HN could be more hospitable to finding these types of deals. Maybe a new section?
For sure. As a seller I think services like FE make a lot of sense. They have relationships with a lot of potential buyers to put your business in front of. Unless you're well connected it's going to be a lot harder to drum up buyers on your own. As a seller what you want is an auction where multiple buyers are competing to buy your business.
