The Relative Value of Domain TLDs Part 1

A big question in valuing domains is how much each TLD is worth relative to each other.

Normally these relative values are taken with respect to the .com value as .com is the most valuable and the TLD for which most data is available on.

Everyone’s approach seems to be to estimate the value of the .com and then discount it by some fixed multiple to suit the TLD it’s on.

I liked the idea of having one algorithm to determine the value of a .com domain and then discounting that value for the TLD it was on but I knew that the idea there could be a fixed discount rate for different TLDs was out of touch with reality.

A tremendous amount theorising and study has gone into this problem and I have heard a number of experts mention different multipliers that they use in their valuations.

For example “a .org is worth 10 times less than a .com”

Fixed discount rates

Thies Lindenthal a far smarter and more qualified data nerd than I am, came up with the best answer to this problem in his study of the domain market.

I didn’t feel the need to repeat such brilliantly done research.

Thies reckons that the TLD pecking order is as follows:

COM > DE > CO.UK > NET > EU > ORG > ES > INFO > BIZ > MOBI

Based on this research .nets are worth only 25% of a .com, .orgs just 16% and .infos a meagre 10%.

This all seems seems reasonable considering our knowledge of domains but this of course doesn’t deal with the relationship between the domain and it’s TLD something Thies by the way acknowledges in his research.

I thought I’d look to the data to see if my hypothesis that there are variable and quantifiably different discount rates for domains with different relationships to their TLD.

DomainMarket.com

I studied the complete public DomainMarket.com portfolio. They have 168,112 domains listed with buy it now prices and 45 of those domains were registered in more than one TLD.

I categorized these domains into:

  1. .coms that had a .net match
  2. Charity related .coms that had a .org match
  3. Non charity related .coms that had a .org match

I then calculated how much more valuable the .com was compared to its .net and .org duplicates by taking an average of the multiple by which the .com was priced higher than the .org/.net

What surprised me was the incredible spread in the data. I was sort of expecting DomainMarket to have an standard conversion rate but that proved not to be the case.

From the 11 domains registered in both .com and .net the average .com was 55 times more valuable than its .net counterpart.

This average was polarised by very large and very small multiples, which discredits the value of the average but is interesting in its own right.

There is no apparent reason why a .com and a .net would target a different audience and thus the multiplier across different domains should remain fairly constant. Yet we do not see this in DomainMarket’s pricing.

For example tasty.com is on sale for $1m yet the .net version will set you back $90k – 11.1 times less than the .com

HappyBirthday.com is also on the selling block for a million dollars yet its .net version is only $30k – 33.3 times less than the .com

And then you get to the other end of the scale with LittleDesigns.com/net priced at $40,000 and $400 respectively, making the .com 100 times more expensive.

There is no obvious reason for these discrepancies which makes this pricing all the more interesting.

Perhaps a more potent learning from the data is in how DomainMarket.com have their .org inventory priced and structured.

There were only 4 domains on which they owned the .com and the .org to which I felt there was no charitable connection in the domain. In this category the .org was 50 times less valuable than the .com

But DomainMarket.com have clearly been buying a lot more .org domains when they suit the TLD. They held 18 domains in both .com and .org where the domain was clearly related to some form of charity. In this category the average .org was only 2.64 times less valuable than the .com.

In fact two domains, childrensvoice.com/.org and charityconcerts.com/.org were actually more expensive as .orgs.

What’s evident in the data is that domains that suit the TLD they’re on are worth more than domains that don’t suit their TLD, an obvious statement it would seem but one not represented with fixed discount rates for different TLDs.

This learning of course could be extended to any other TLD, you would expect that business related .bizs be of greater value than charity related .bizs and British .co.uks more valuable than French related .co.uks.

I took this data driven learning and then tried to represent it in the data.

It’s incredibly hard to represent human behaviour with data but I found a fantastic data point that represented exactly what I wanted.

The percentage of .orgs in the top 100 results in a Google search turns out to be an excellent measure of how charitable that search is.

When you think about it, it seems so obvious.

Charity related sites tend to choose .orgs and Google tend to show more charity related sites for charity related searches – so if you reverse this and do a search you can measure how charitable a search is by counting the number of .orgs in the top 100 results.

In my study of 12,000 Google searches it turned out that on average 10.13% of the top 100 results for a search query were .orgs so any significant increase from this would represent that a search term was charity related.

This point is more extensively argued in my article documenting this measure of how well suited a domain is to it’s TLD.

Of course this data point can then be used to determine how British (% of .co.uks) a domain is or how businessy a domain is (% of .bizs).

When you combine this data point with the classic fixed discount rates you can create a variable discount rate for each TLD based on the number of occurrences of that TLD for the domain in question in Google.

This naturally leads to far more accurate domain valuations and drastically improves the potency of automatic domain valuations for non .com domains.

For most TLDs a fixed discount rate works fine, in part 2 of this series I will provide fixed discount rates for dozens of TLDs based on hard data and not human speculation.