IT.COM

analysis Is Domain Name Investing Profitable?

Spaceship Spaceship
Watch

Bob Hawkes

Top Member
NameTalent.com
Impact
40,786
In a recent NameTalent post an analysis is made of how profitable domain name investment is on average. The starting premise is:

"A domain investment will be profitable if the probability that the domain name sells in any given year multiplied by the net profit on the sale of that domain name is more than your annual holding costs associated with the domain name."

For example, if your net profit (i.e. after costs considered) on a domain name is expected to be $1500 and your probability of selling in any one year is 1/100, then the investment is profitable if your annual holding costs are less than $15.

I used sales data and number of domains listed for sale to estimate the probability of a sale. I did this for all extensions together (and data for the last year), but the market is so dominated by .com that .com alone numbers would not be much different.

A correction factor must be applied for the fraction of total sales that NameBio stats represent. I used a factor of 5 (i.e. 20% of total sales are on NameBio). With this factor it suggests that the chance of any one domain name selling in one year is 1/73. So if your portfolio is 350 you would sell about 5 domain names per year.

In the initial work I used the NameBio average selling price. There are arguments that NameBio may under-estimate average prices (wholesale price bias), or that by including the super premium sales it may over-estimate the value expected by a typical domain name investor and we should use the median price.

For costs I assumed commissions at 10%, other selling costs, and a cost for the time the domainer puts into handling the name. Note that while I take into account most types of costs, things like a portion of Efty memberships or web hosting costs are included with renewal fees in the annual costs, rather than applied in the net profit line. Clearly it could be done either way, as long as consistent.

I applied this relationship using several different models as shown in the table of results below.
ProfitModels.png

The first (yellow) average model is my best estimate. I assume that you buy the domain at $70, sell at NameBio average, pay 10% commission and $25 other costs (this includes a portfolio averaged cost for things like legal, financial, art, evaluation fees, etc.). Arbitrarily I assume that you should get $100 for the time you put into handling acquiring, promoting, maintaining and selling (maybe 4 hr at $25 per hour). You also gain by any profit in bottom line. I assumed a pretty aggressive $10 annual cost assuming you have some sort of volume discount on renewal and have a big enough portfolio that your web costs are small (or you use only marketplaces). The final row in the table gives you the net ratio - e.g. in this situation it suggests profitability with a 1.31 ratio (below 1 means you lose money, on average).

Under median I kept most things the same but I assume that you sell at the median, rather than mean, price. The NameBio stats don't allow you, anymore, to calculate it for the entire large dataset, but when I do it individually each day it is usually in the $275 to $400 range. I used the upper value. Here we see things are not profitable.

The low model assumes you buy at reg fee, sell at median price, reduce commission to 9% and keep other costs and the cost for your time at $0. Even at this not quite profitable.

Premium is meant to apply to a premium ngTLD or premium country code situation. Assume somewhat arbitrarily you buy at $200 with $100 annual (mainly renewal) fee. Here assume that you sell at about double the average price, $2500.

Some of us start in domain investing by hand registering some inexpensive ngTLDs. The cheap model simulates this. I assume that you buy at $1 and manage (perhaps with multi-year discounts) to keep renewal at $6. But you only sell at $250 and don't value your time. The probability of sale should probably be reduced, which would make the situation even less favourable.

The optimist and pessimist models are meant to look at what might be the bounds of the likely models. Optimist assumes that really only 10% of sales are in NameBio, you sell at about double NameBio average, you sell directly without commissions and don't add much for your time. The result is very optimistic with a ratio of 8 :xf.smile:. For pessimistic, well pretty pessimistic :xf.frown:.

You can read the full post at NameTalent at this link:

https://nametalent.com/2018/10/is-domain-name-investing-profitable/

Keep in mind that these are probabilistic arguments for the domain business as a whole. Clearly many successful investors will do better than this, while a number will do worse. Also, I totally accept that some of the numbers (like the NameBio correction factor) are uncertain.

Thanks for reading!

Bob
 
31
•••
The views expressed on this page by users and staff are their own, not those of NamePros.
@MetBob thanks for your in-depth analysis...
I am sure it took lot of your time..you are passionate about the stats I assume and have lot of patience...
Lucky us... getting some good insights.

Thanks
Ravi
 
5
•••
Thanks for your kind comments @vravis9. I like to do analysis so happy if I can contribute to our domain community in this way.

I meant to point out to anyone if they would like me to run numbers with a different set of numbers I am happy to do so. Just tell me which of the following numbers you would like changed and to which values to see results for.

Number For Sale (I assumed 30 million)

Factor (Real Sales/NameBio)

Acquisition Price

Sale Price

Commission (tell me % you would like used)

Other Costs

Your Time Costs

Annual Cost

I also would point out that I did not put in a factor for costs of borrowing money to invests in domain names (or forego investment income on funds you could have otherwise invested0. One way to do this would be to add to annual cost line. Like if you borrowed $70 for a domain name acquisition at a 5% interest rate then there is in essence an extra $3.50 per year cost. I may work this into a future model.

I plan to refine the model over time, so welcome suggestions, comments, corrections.

-Bob
 
5
•••
Hi
so, how has it helped you, in your quest to profit from domaining?

imo….
 
1
•••
Thanks.

For namebio, it might make sense to exclude all xxx,xxx sales, all low xxx sales, especially for LLLL.COMs as those are pretty much inventory purchases and beneficiaries are registrars (in fact all expired domain auction and dropcatch auction results should be excluded)

This will bring your probability of sale under 1/100.
 
5
•••
To base probability estimations on Namebio sales count (or any other aggregated numbers) is at best inefficient.
 
1
•••
@MetBob Bob the analysis is already comprehensive and covered all possible factors...
Coming to borrowing money to invest in domain purchases is basic information that particular individual should include wherever it is applicable...

I want to know if it is possible to see the total picture of above mentioned stats to respective to individual sales platform...
It doesn't make sense if we check for all marketplaces...
I am talking about major marketplaces like Sedo..Afternic..Namejet and GoDaddy..since these get lot of sales and some are more popular with enduser sales...and some reseller sales...
I got to know about this in in-depth from the article you (@MetBob ) shared previously.. about NZ Domainer analysis method...

I think if we can get this data...it will definitely give a greater look into basic / low ball sales....and which marketplace is getting these type of sales and how frequently...

This can be easily related to or can be a guide for beginners...

Thanks
Ravi
 
3
•••
Thanks for your excellent question @biggie.

In one sense I would say it has not helped directly with being profitable in my own domining (maybe the opposite, I should have put my time into trying to sell my own names!).:xf.wink:. It is a look not at my own, or anyone else's, portfolio, but rather statistically.

However I think indirectly considering and developing the model is helpful to me, and I hope others, in the following ways.
  1. It helps one be realistic with expectations. The ratio of sold/for sale in any one year is small, maybe 1 in 75.
  2. It helps me be logical in my consideration of potential acquisitions. What really is my probability of selling this in one year, and what is a realistic sales price? If I put them together, and compare to my annual costs, does this acquisition make sense?
  3. By seeing how important costs are, i think it helps remind me to keep costs as low as possible but only if I can do so without reducing sales probability.
  4. To some degree it provides guidance on different types of domain investments (although really the next phase of my model that will adjust sales probability and price to extension/type to do that).
  5. I think it helped me, at least, think about how much my time is worth re each domain name. I think we often overlook this factor.
  6. It helps inform a business model for domain investing. This can be important if you are needing to convince others that you have a workable plan going forward.
Bob
 
Last edited:
8
•••
To base probability estimations on Namebio sales count (or any other aggregated numbers) is at best inefficient.

Can you provide a bit of explanation by what you mean by inefficient? I think, as the full report discusses, we need to be alert to biases introduced by the fact that they only sample portions of the total sales. But I am not sure what you mean by inefficient?
 
2
•••
Thanks.

For namebio, it might make sense to exclude all xxx,xxx sales, all low xxx sales, especially for LLLL.COMs as those are pretty much inventory purchases and beneficiaries are registrars (in fact all expired domain auction and dropcatch auction results should be excluded)

This will bring your probability of sale under 1/100.

Thanks for interesting valuable suggestions @Recons.Com. Just to make sure I don't misinterpret, is this what you suggest?
  1. Exclude the 6 figure sales as not representative of (most) individual domain investors
  2. Exclude low $$$ sales (maybe below $200?) since those are wholesale domainer to domainer sales?
  3. From what remains after these two exclusions calculate a new sales number and average price to use in the calculations
Thanks,

Bob
 
Last edited:
2
•••
@MetBob Bob the analysis is already comprehensive and covered all possible factors...
Coming to borrowing money to invest in domain purchases is basic information that particular individual should include wherever it is applicable...

I want to know if it is possible to see the total picture of above mentioned stats to respective to individual sales platform...
It doesn't make sense if we check for all marketplaces...
I am talking about major marketplaces like Sedo..Afternic..Namejet and GoDaddy..since these get lot of sales and some are more popular with enduser sales...and some reseller sales...
I got to know about this in in-depth from the article you (@MetBob ) shared previously.. about NZ Domainer analysis method...

I think if we can get this data...it will definitely give a greater look into basic / low ball sales....and which marketplace is getting these type of sales and how frequently...

This can be easily related to or can be a guide for beginners...

Thanks
Ravi

Hi Ravi,

Thanks for your comments.

I did consider whether to include only certain sales venues, since clearly some have more wholesale and some mainly retail, and some are registry related. However, in the end did not, largely because of the desire to try to see if taken in entirety, how profitable, or not, is domain investing. Practically most of us make a mix of money from sales to end users and to other domain investors.

While it would be technically easy to include or exclude different marketplaces, I am not sure that it is the right thing to do unless you want to answer a slightly different question on only retail sales.

Bob
 
Last edited:
2
•••
Thanks for interesting valuable suggestions @Recons.Com. Just to make sure I don't misinterpret, is this what you suggest?
  1. Exclude the 6 figure sales as not representative of (most) individual domain investors
  2. Exclude low $$$ sales (maybe below $200?) since those are wholesale domainer to domainer sales?
  3. From what remains after these two exclusions calculate a new sales number and average price to use in the calculations
Thanks,

Bob

Yes, that is correct. Maybe, also exclude all expiring/dropcatch results as well, if there is such possibility, as those definitely are not investor sales and not relevant. I would also exclude all LLLL.com sales below $500.

I anticipate that with all these done, your average sale probability might go from 1/73 to closer to 1/150.

Average sale price might go up even with exclusion of xxx,xxx or even anything above $40,000 really.
 
2
•••
Thanks. I don't know of way to technically exclude expiring/dropcatch except by excluding a couple of venues that specialize in that. Any below $100 are of course already excluded since NameBio do not report them publicly. In the same way some registry but not all registry can be readily excluded.

Yes, that is correct. Maybe, also exclude all expiring/dropcatch results as well, if there is such possibility, as those definitely are not investor sales and not relevant. I would also exclude all LLLL.com sales below $500.

I anticipate that with all these done, your average sale probability might go from 1/73 to closer to 1/150.

Average sale price might go up even with exclusion of xxx,xxx or even anything above $40,000 really.

I suspect that you are right that the probability will go down to something like that but when the (probably) increased price goes up the net result may be less than at first expected on the overall ratio.

One approach I considered was to use just Sedo data but the correction factor to the full dataset seemed so large I feared throwing out that much.

Lots to think about. Thanks again for your input.

Bob
 
1
•••
Does godaddy report anything besides its auctions? If no, then you can safely exclude GD (but keep afternic), NJ as those are pretty much reseller trades. While there is some profit there, you can just exclude that segment and basically do your analysis on end user sales, which is what most investors area of interest anyway.
 
3
•••
Does godaddy report anything besides its auctions? If no, then you can safely exclude GD (but keep afternic), NJ as those are pretty much reseller trades. While there is some profit there, you can just exclude that segment and basically do your analysis on end user sales, which is what most investors area of interest anyway.

My understanding is that it is only the expired auctions on GoDaddy, so like you say they could readily be included (it seems there are some entries that are not from expired, but not many). Here is a link to Micheal's reply to my question in a different thread re what is included.

https://www.namepros.com/threads/namebio-database-sources.1073148/#post-6642007

Although I can keep Afternic, almost none of their data (from recent years) is in NameBio since they stopped reporting. Thanks again for suggestions.

Bob
 
1
•••
Thanks @MetBob for this kind information
 
1
•••
Can you provide a bit of explanation by what you mean by inefficient? I think, as the full report discusses, we need to be alert to biases introduced by the fact that they only sample portions of the total sales. But I am not sure what you mean by inefficient?

There are X domains registered and Y sales per year (let's assume we know the exact number of sales), does that mean that any domain has a Y/X probability to sell per year ? Can I reg sga67s89-sadsa-d7.biz and expect it to have, for instance, 1% probability to sell per year ?
 
2
•••
There are X domains registered and Y sales per year (let's assume we know the exact number of sales), does that mean that any domain has a Y/X probability to sell per year ? Can I reg sga67s89-sadsa-d7.biz and expect it to have, for instance, 1% probability to sell per year ?

No of course not :xf.wink:.

What the statistics apply to are the universe of names that someone registered and is trying to SELL. It does not mean that any individual domain name has that probability. It does mean, however, that the industry as a whole, if we accept definition of industry as being all the domain names for sale, is represented by the probability (within its uncertainty).

It means that if you are an average seller, i.e. not more intuitive and skilled than the average, but also not less than the average, the probability, in a long enough time period, would be relevant.

Of course in most things most people think they are better than average! Maybe I will conduct a NPs poll to ask that :xf.smile:

Bob
 
1
•••
It does not mean that any individual domain name has that probability

But this is what you are using in your formula.
 
0
•••
But this is what you are using in your formula.

The formula gives the statistical probability (to the degree that the numbers are known) that would statistically apply. It is like saying the probability of a temperature on a certain date having some value. It does not mean the temperature that day will be that, or that the statistics are wrong when a day has a different temperature.

I'm not sure if you read the entire article at the link. The closing paragraph (and elsewhere) stresses the statistical nature. I bolded a sentence for emphasis below.

While the overall picture presented here may be discouraging, we should also keep in mind that a number of domain investors have been successful year after year over an extended period in the business. As with any statistical argument, they show a general picture, not what applies to any one person. Outside domain investing most startup businesses fail within a decade. That does not mean that your startup will necessarily fail. Being quantitative in your business plan model will both force you to have realistic expectations and to make logical domain investing choices.
Also, in my original post in my thread I made a similar remark (again I added bold here for emphasis).

Keep in mind that these are probabilistic arguments for the domain business as a whole. Clearly many successful investors will do better than this, while a number will do worse. Also, I totally accept that some of the numbers (like the NameBio correction factor) are uncertain.
Bob
 
Last edited:
1
•••
What you are doing is similar to this : You get the full database of NBA games and notice that the Home team has some advantage (over 100,000 matches). Let's say the Home team wins 55% of the time and the Away team only 45% of the time.

Now there is an NBA game later tonight, so without knowing which teams are playing, you estimate that the Home team has a 55% probability to win (which is correct, absent any further information). But someone who checked the fixtures and knows which teams are playing (could even be a David vs Goliath case) is in a better position to estimate the probabilities of the outcome. Because they have this extra piece of information.

That's why I say your method is inefficient.

In our case, domains, this extra piece of information is the domain itself.

The formula you posted treats the probability of a sale in a uniform manner over all domains (ie. as a fixed number). It is inefficient. It is way more efficient to estimate the probability of a sale for each domain individually. Moreover, this probability is dependent on the BIN or asking price (obviously).

There are other inaccuracies in your post too. Sorry to have to say that, because I know your intentions are the best. And I know research and publishing takes a lot of time too. But some people may take everything at face value and act on these findings.
 
3
•••
Thanks for laying out your reasoning of what you meant by inefficient @Asfas1000. I agree that this is exactly like your NBA home team example.

I think everyone on here agrees that the domain name matters in the probability for selling (and price). Now the formula itself (but not the numbers) can apply to an individual domain name as I explain in the OP:

For example, if your net profit (i.e. after costs considered) on a domain name is expected to be $1500 and your probability of selling in any one year is 1/100, then the investment is profitable if your annual holding costs are less than $15.
Yes, absolutely, if you can from experience, other information, whatever, estimate the probability for each individual domain in your portfolio, and estimate the expected price, then you can decide on whether each domain in your portfolio makes sense as I did in the example above.

The article deals with how, on average, people do. The article at the link starts with the following where I lay out why I went through the effort of developing a model. I was concerned that too many jump into domaining unrealistically expecting easy big profits. I wanted to see what the average numbers say.

I am concerned when I read posts from new domain investors saying something like “I registered 20 domain names last month and they have not sold. What am I doing wrong?” or “I just quit my job and plan to make easy money flipping domain names.” Some of the books written on domain investing perpetuate the myth that it is fast and easy to make big returns in domain investing. Let’s take a look at what the numbers really say about how profitable domain name investment is on average.

The purpose is foremost to look overall at the profitability of the industry, and then I looked at different models within that partly to set bounds between say the most optimistic and the most pessimistic models.

I am working on a future post on exactly the idea of applying the formula to individual names in your portfolio. However, as background to that I think it is valuable to look at the industry wide numbers. Otherwise, I think it would be too easy to fool ourselves into unrealistic expectations, say I have 1 chance in 2 to sell this domain this year for $10,000. I think is is important to see how rare that would be. I also of course offered in the thread if anyone wants a set of numbers they feel apply to their situation run to do that.

You mention inaccuracies without being specific. If you share anything I am glad to respond. I would point out that the linked article takes some time in pointing out some of the uncertainties, and also that things like investment costs on borrowed money are not included.

Bob
 
Last edited:
1
•••
it's more complex

as the profitability is not only gained by a sale
but the domain meanwhile may have income based on its traffic

additionally the domain sold must pay for the other thousands unsold domain renewals

the question is:
can you sell a sold domain high priced enough
to make it work as whole
 
Last edited:
2
•••
as the profitability is not only gained by a sale
but the domain meanwhile may have income based on its traffic

Thanks @frank-germany. I should have stated that I assumed no parking or similar revenue. If you do expect revenue, a simple way to enter in model is simply subtract from annual costs. That is if your renewal is $9 other annual costs $1 but expect to earn $6 parking, the annual effective cost is $4.

The model does already handle the maintence cost of the rest of your portfolio.

Bob
 
3
•••
0
•••
Back