Valid point! But still, getting an idea on what to retail for, not generalizing it, would help.
In that case, we should segregate at our end and see that we are looking at the right set of data.
I am talking about a use case where you are putting it for BIN. You need to put up a price and for that, you need a reference, isn't it?
The easiest way to find the retail sales after you select your filters is to just sort by price descending instead of worrying about the venues. You'll miss a lot of random sales from domainers/brokers only looking at certain venues. Although you're right that some venues are predominantly one type or the other.
Pricing is different from appraising though. You can fairly easily do a wholesale appraisal because there's an established market with lots of data. I still don't think it's possible to do a retail appraisal, if you want proof of that watch the older DomainSherpas where they used to play the "Name That Price Game" and guess what price a retail sale went for. The guesses were all over the place. You can obviously come up with a retail price to put on your names, but it's a spectrum not a science.
There's not really a right answer when it comes to retail pricing, it all just boils down to this: the higher you price it the lower the chance it has of selling and the longer it will take. It's about finding a balance that works for you. If you owned Cars.com and you priced it at $1k it would be sold in seconds, if you priced it at $2m it would be sold in a matter of weeks, if you priced it at $15m it would probably take several years, and if you priced it at $10 billion it would never sell. The same applies to any domain.
It depends a lot on your own personal strategy and cash flow needs. Lets say, for example, that you mostly used comps from BuyDomains to price your names. Well they own a boatload of domains and have a churn-and-burn strategy, so they're pricing low and making up for it on volume. They can do this both because of the size of their portfolio and because they're exceptionally good at replacing inventory on the cheap. But if you own 100 names that are solid quality, and you have trouble replacing them, following their lead on pricing would be a mistake. You would want to emulate someone who has a similar model to you.
I think if you look at enough retail data you'll just end up pricing everything in a fairly standard way. Low quality names that are easy to replace in the $2k - $5k range, good names in the $5k to $25k range, and great names left as "make offer". And where you lean in the range would depend on your own strategy.
Yes, I guess if you're trying to price for high velocity, and most comps are in the $2k range and a couple of outliers are in the low five-figures, that would give you a pretty decent idea of where to price it at. Or if you're feeling spicy and want to shoot the moon you could aim towards those outliers. But again I think most people will just end up towards that standard anyway.
I think retail comps are most useful in negotiations, not pricing. Although it would be a good idea to look at the comps first to make sure you can justify your price if you get called out on it.