Yesterday Theo Develegas from Acroplex wrote a blog post about how some reported sales from GoDaddy are bogus and never actually completed. Rather than writing a long reply in the comments section of his blog I thought I would cover the topic in an article. I will discuss how data is collected from each venue and the issues faced with each one. I will also explain why his proposed solution of verifying ownership changes through WHOIS isn’t practical or reliable.
Let me start off by saying that the domain industry isn’t transparent like residential Real Estate where a sale price is a matter of public record. Most venues are not willing to share their sales data and those that are usually do so sparingly. It is very important for the domain industry to have a searchable database of sold domains, it not only helps determine the value of other domains but also helps to understand the state of the market. We must have this data, so if the venues won’t cooperate we need to do our best to collect it ourselves.
Data Collection & Issues
GoDaddy
I’m starting with GoDaddy because this is the venue Theo discusses in his article. GoDaddy goes out of its way to make it difficult to track sales. As soon as an auction ends it is removed from your watch list unless you were actively bidding on it, so unless you were participating in the auction or watching when it closed you can’t see the final price. There are two issues with GoDaddy that arise after the auction closes.
The main issue is that it auctions domains that are still able to be renewed by the owner. After an auction closes we record the final price, but the current owner has up to seven days to transfer the domain to another registrar or renew it and invalidate the auction. We don’t hear about it when this happens unless the winning bidder contacts us, which does happen but not often enough. The sale remains in our database even though it never actually completed.
The second issue is that of non-paying bidders. After an auction closes, if the winning bidder doesn’t pay the domain is offered to the runner up with all of the winning bidder’s bids removed. In this case the sale completed, but at a lower price than we recorded in our database, potentially much lower depending on how the bidding played out.
Flippa
Flippa is our favorite venue as far as transparency goes. Its director of domains, Kevin Fink, is rabid about data being accurate. Not only do they provide us with a feed of all closed domain auctions, Kevin follows up on sales to make sure they completed and notifies us when they don’t so we can remove them from the database.
A Flippa auction that closed yesterday and hasn’t completed yet might never complete and could be wrong, but eventually Kevin will notify us and it will be removed. So after a few weeks the data is accurate.
Sedo
All data from Sedo is reported by the company only once the sale has completed, so this data is always accurate. They have an RSS feed of completed sales that updates frequently, and then they send a spreadsheet once a week of completed sales. We don’t track their auctions on a daily basis so there is no concern about non-paying bidders or sellers who refuse to transfer the domain.
NameJet
NameJet provides a spreadsheet of completed sales once per month, but it is a fraction of their total sales. For NameJet we have to place back orders on every auction we want to track. We place in excess of 50,000 back orders per year now. If the winning bidder doesn’t pay the runner up can take a stab at it, so sometimes the sales actually completed at a lower price than we recorded.
NameJet has a more difficult verification process than GoDaddy so nonpayment isn’t a serious problem. However, NameJet wins can be clawed back even after the domain was transferred to the winner and we wouldn’t hear about this, so the sale would stay in our database despite being reversed.
DropCatch
This venue doesn’t report to us, we have to watch the auctions as they close, so the data isn’t verified. That said, they re-auction domains when the winning bidder doesn’t pay, so it is easy for us to spot when an auction didn’t complete and remove the original record from our database. If a domain sells twice at this venue within a week it means the original auction fell through.
BuyDomains
All sales are reported by the company once completed so the data is always accurate.
Fabulous & 4.CN
Both companies provide a feed of sales, but I’m not sure if they are put in the feed when they close or when they complete. If anyone knows for sure please let me know in the comments.
Why Ownership Verification Wouldn’t Work
First and foremost, we collect on average about 5,000 sales per month. Manually checking the WHOIS history on that many domains would be a full time job and we simply don’t have the resources to dedicate to that. It might be worth it (emphasis on might) if ownership verification actually confirmed our data, but for the most part it doesn’t.
Imagine an auction closed at GoDaddy, and then eight days later the WHOIS changed from GoDaddy Redemption Services to a new owner who wasn’t in the WHOIS before it expired. What does that really tell us?
Let’s say Bidder 12 won the auction, do we know he is the person who is now in the WHOIS? Maybe he didn’t pay and the runner up is now in the WHOIS, but he paid a lower price than we recorded. Theo suggests we could take that as confirmation that our data is correct because it wasn’t renewed or transferred out, but in reality we can’t have any more confidence in our data than if we didn’t even bother trying to verify it.
Take another scenario where the auction closes, but then the previous owner renews it or transfers out and immediately puts it under privacy. We would have no idea if it was renewed, if the winning bidder paid and took ownership, or if any of the other bidders took it at a lower price after the winner didn’t pay. Do we remove the sale or assume it completed? We have a 50% chance of making the wrong move.
The only time ownership verification would actually give us a definitive answer is when the owner renews and keeps the WHOIS public. In that case we could confidently remove the sale from our database, but this happens infrequently enough that it isn’t worth dedicating a full-time employee to detect when this happens.
We’re Doing the Best We Can
There isn’t a good alternative to make sure the data is perfect. We can’t leave out all GoDaddy sales and NameJet sales that aren’t reported by the company because they can’t be verified. We would be missing half the market just to make sure a tiny percentage of uncompleted sales don’t make it into the database.
So we’ll just have to live with knowing that some sales never completed, or completed at lower prices, until more venues start cooperating with us. It’s better to be 97% right than be completely in the dark. I’m a perfectionist so I don’t say that lightly.
Michael Sumner is the CEO of NameBio.com. Previously he was the lead developer at State Ventures which owns and operates geo domains such as OceanCity.com and Maryland.com. Michael was also the co-founder of DN Media, a company that was involved in seven figures worth of domain name transactions.
Michael, the situation with GoDaddy is unique and creates many challenges, on which I expanded in two recent posts.
In the latest one, I covered the issue of how such “non sales” that aren’t unverified, are covered by industry blogs – sometimes before, during and after the ending of the auctions. The end result, is a growing list of sales that did not complete.
Data accuracy is important in every industry, particularly that of financial nature. I think it’s better to not report a sale that has no confirmed numbers attached to it, than to declare a sale as complete, with a figure that is not true.
Unfortunately, confirming the validity of data is important and it comes with the territory. Imagine if any numbers were thrown out there for any domain whether it sold or not.
This is squarely a GoDaddy issue, as you accurately broke down the process of other venues.
Unfortunately it isn’t just a GoDaddy issue. As mentioned, NameJet sales can get clawed back even after the domain is transferred to the auction winner. This is the exact same situation as GoDaddy with the domain getting renewed after the auction is paid for and the sale being reported by industry sites even though it was ultimately reversed.
In my view this is the same issue as non-paying bidders, which is a problem both for pending delete and pre-release auctions and happens at all venues. In this case it isn’t as bad because the sale does complete, but the price that gets reported is higher than what the winner actually paid.
I don’t think what is happening is quite as dramatic as intentionally reporting anything as a sale whether it happened or not. In the case of GoDaddy auctions getting renewed at the last minute, the price reported is what the market valued the domain at. I think non-paying bidders is a more serious situation because then the prices are often very inflated and not a reflection of the market.
Either way, we don’t have the resources to spend a few thousand dollars a month to delete a few dozen sales out of several thousand. Hopefully more venues will cooperate with us like Flippa does, or bidders will report to us when an auction was reversed so we can remove the sale.
While Namejet can clawback a sale, in my experience, it happens much more at Godaddy than anywhere else. For me, and maybe it is just me, it’s 10x more likely to happen at Godaddy.
Validation must come at the source, e.g. GoDaddy, just like Sedo releases data on completed sales.
On the subject of how non-completed sales are touted as “sales”, often the moment they occurred and before even payment is rendered, I don’t believe personally that the closing numbers qualify as indicators of value. I plan to expand on this in a future post.
I agree that the task of checking every sale for accuracy might be herculean, and thus is the challenge. I propose that you reach out to GoDaddy (and NameJet) so that “final” sales lists can be fed to the database.
Most venues consider their sales data proprietary because it can be used for business intelligence and appraisal algorithms / price suggestion tools. If they report, they only report the high dollar ones for the free publicity, but there are thousands of sub $2k sales every month that don’t make headlines but are important comps which they have no interest in sharing. As you mentioned, the only way perfect data is feasible is if the venues cooperate, but I don’t see that happening any time soon.
I agree that a closing price is not an indicator of value, but only in the case where the winning bidder doesn’t pay and didn’t have any intention of paying. If it was a legit auction that only didn’t complete because the previous owner renewed, the closing price would have otherwise been the official sale price. So in that case I think it is a perfect reflection of what the market would support. Looking forward to your post on the subject.
These sales are for comps. So why bitch about less than 1% off all transactions that might not be correct. This isn’t balancing of my bank account or a payment on my house. These are merely results that can be used to determine the value of another domain. How does this even effect anyone financially? Some people spend way too much time sitting around trying to figure out something else to complain about. In my opinion people should be thanking Namebio for the incredible database of past sales and the auction houses for providing liquidity. Unless a domain investor posts every sale, who it was sold to and the price as well as every purchase and its price on their blog, I can’t take anything they about companies needing to show transparency, seriously.
Shane – You need to go back and read exactly what the scope of my analysis was about, so spare me the football coach commentary about “bitching” etc. Those that know and care about accuracy of information do some due diligence.
My post wasn’t about NameBio, it was about the extended effect of GoDaddy’s management of domains on auction. Sales that never completed are listed as sold when one searches for the domains. I’m not the only one that identified the problem, so it seems you’re barking up the wrong tree, again.
If you’re happy with posting lists of “XYZ domain sold for $123” that’s fine with me. My role is to perform an analysis and to offer suggestions, and I will continue this approach when something doesn’t work as it should.
Great and informative article. Not shocking that there is a lack of transparency with some sales.
Michael,
That’s great you reveal how you collect data, interesting.
My feeling is anything you can do to have more accurate results without expenses skyrocket should be done. Maybe having a script that automatically checks the WHOIS of GoDaddy/NameJet auctions before and a pair of weeks after the closing date will help NameBio know confirm or no the sale.
It’s was my understanding that it was the same people behind Namebio and Whoisology so you should have an easy access to a WHOIS API at no/low cost.
In a side note, congrats, as the site has really improved since you are behind.
I think NameBio makes a great job collecting all these sales data. It’s a valuable tool for domainers worldwide!