Flippa has this new blog post about various ways to value sites.
I don't agree with much of it and have argued in the past that examining multiples is worthless unless it's multiples of net income and you don't see net income in listings, that's something you have to work out manually yourself by weeding out all the bullsh*t in the description.
But comparative valuations have got to be a good way, right? It's what I used when creating my own valuation engine.
Not so. I argue now that the fundamentals have changed.
When I wrote the valuation article for Sitepoint I collected data on hundreds of sites, normalized their earnings etc., before calculating what multiples various categories of sites sold for. It worked then.
It don't work no more.
One of the problems then was that there weren't a great number of places to take samples from and without enough samples you can't draw reliable conclusions. Brokers wouldn't share data of private sales and sale prices, they didn't need to. The only source for pricing was Flippa.
That's still the case, the only source for pricing is Flippa. But the quality of data has deteriorated. NPB have become, I feel, an endemic problem. Just because a site appears to have been sold in Flippa doesn't mean it's sold. There's a good chance the buyer backed out. Using it and the sale price in your calculation will, therefore, include highly overpriced and spurious bids that were not backed by any real buyer intent.
Someone did a blog post recently and drew some conclusions from the "final selling prices" of various Flippa listings. As I pointed out in there, one of his examples that supposedly sold for $200K was "sold" a month later by the same seller for $40K. That happens a lot.
Can we brainstorm for ideas? How can we clean the data, use a different data source or otherwise improve the reliability for calculation on comparative valuations?