# strategyWhat Are The Odds? Applied Probability For Domain Investing

Many skills can help you make better decisions as a domain name investor. One of them is quantitative decision making. While probability can never firmly tell you what to do, it can help make better decisions in some cases.

Applied probability can be used to help answer questions such as how many domains will I probably sell this year, should this domain name be renewed, and whether it is better to renew names in advance of a forthcoming price increase.

Basic Ideas

A probability is a numerical estimate of the chance of something happening. For example, if you roll a normal dice, unless it is rigged in some way, the probability of any particular number being rolled 1 chance in 6.

Normally we express probability on a scale from 0 to 1, with 0 meaning no chance that result will happen, and 1 meaning it is sure to happen. In the dice example, the probability of each number would be about 0.167.

Sometimes probabilities are expressed as percentages, from 0 to 100%, rather than from 0 to 1.

If we go back to the dice example, if we roll it twice, the odds of getting 2 followed by another 2 will be about (1/6)*(1/6), or one chance in 36. That is because the events, for a fair dice, are independent.

It is important to realize that probabilities are not always independent though. For example, the probability of a domain name in a certain sector selling, and the probability of a high sales price, are probably related. If demand for a niche or sector goes up, the chance of a name selling and its price, both increase.

A probability value may change with time. For example, the probability of a metaverse domain name selling 10 years ago was probably less than this year. Some sectors that were hot a few years ago are much less active in domain name sales now.

If you want to read more about probability, the Wikipedia entry on probability is well written.

The Sell-Through Rate, STR

The sell-through rate (STR). STR is simply the number of sales during a 12 month period divided by the average number of names in your portfolio during that year.

If you have a substantial portfolio, and have been investing in domain names for a number of years, you can calculate your annual sell-through rate (STR) each year.

Many include only sales above some price point in calculating their STR.

EXAMPLE â€“ STR Calculation:
Letâ€™s say someone sets \$600 as minimum price for their STR calculation. If they sold 5 domain names above that price during the 12 month period, from a portfolio that ranged from 400 to 600 domain names during the year, averaging 500 names, then the annual STR = 5/500 or 0.01 which is 1%.​

The STR is not a probability, but the probability of selling a domain name during a year is expected to be similar to your annual STR, unless something has changed in your domain investing approach.

As a new investor, you will not have an established personal STR from records from previous years. One approach is to use the domain name selling universe, making the rough assumption that your performance might be equal to the average of all names listed for sale. That is relatively easy to calculate as shown below.

EXAMPLE â€“ Industry-Wide STR:
Letâ€™s look just at `.com` domain names and the last 5 years to be representative of current conditions. I am going to set the minimum price at \$1000 and the maximum at \$25,000. See below for a justification on using a maximum.​
With the NameBio interface you can readily find the number of sales over the 60-month period, 75,400, and the average price, \$3061. But not all sales are in NameBio, since sales at several popular venues are not generally included. If we assume that 25% of the sales of \$1000 plus are in NameBio, we can apply an adjustment factor of 4x, making an estimate of 67,860 sales per year.​
Using Dofo Advanced search, setting to only `.com` and BIN prices of \$1000 to \$25,000, there were 11,936,067 domain names listed for sale. But not every domain for sale gets in Dofo listings, and also there are some for sale at make offer, or higher prices, that will end up selling within our price range. On the other hand, there are some old listings still on marketplaces that are not truly for sale. Considering all these factors, I applied a net 1.3x correction factor, obtaining 15,516,887 domain names for sale.​
Now we can calculate the industry-wide annual `.com` STR for sales of \$1000 to \$25,000 = 67,860 / 15,516,887 = 0.0044 = 0.44%. The average sales price was \$3061.

This does not mean that will be your STR. Perhaps you have acquired very strong names, or use effective promotion of your names, and achieve a higher STR. On the other hand, many starting out in domain investing probably have a lower STR than the average. Your personal STR may be higher or lower than the industry average.

Why did I introduce the upper cutoff of \$25,000? It has almost no impact at all on the STR, but the cutoff upper price does impact the average price significantly. Those 6-figure and up sales have a huge impact on the average price. But most domain investors, particularly those just starting out, may never have a sale in 6-figures. In fact, many will not even have domain names priced above \$25,000, so I thought it best to introduce an upper cutoff price to obtain a more reasonable average sales price.

How Many Domains Will I Probably Sell This Year?

Some promote domain name investing as easy and fast. This leads to unreasonable expectations that names will sell for good return on investment quickly. The reality, for almost all domain name investors, is far different.

Letâ€™s look at it numerically. If you have, say 50 domain names, how many sales can you expect in your first year? It is a pretty easy calculation, simply the probability that any one name sells during a 12 year period times the size of your portfolio.

Since you donâ€™t have an established personal STR yet, I will assume that you are â€˜averageâ€™ and the probability of one of your names selling is equal to the industry-wide STR calculated above, 0.44%.

EXAMPLE How Many Sales From 50 Name Portfolio?
Number of sales in 12 months = probability of sale for one name x number of domain names in portfolio.​
Number of annual sales = 0.0044 x 50 = 0.21​
This means an investor with a 50 name â€˜averageâ€™ `.com` portfolio will have a bit more than 1 chance in 5 of selling a name at the end of one full year, or will, on average, sell one name in a 5 year period.​

Let me stress again this is the industry-wide average number. If there is an important message, it is that you must strive to be better than average to find success in domain name investing. Also, unless you are very lucky, a lot of patience will be needed. It is hard waiting those many months before your first sale.

It should be stressed that while it is interesting to see how your STR compares with the industry average, the true measure of success if whether you are profitable. Some have a personal STR well below the average, but sell at great prices and are profitable. Others will sell at a higher rate than the industry average STR, but at prices, or with acquisition costs, that make them overall unprofitable.

Should I Renew This Domain Name?

We all face the issue of which domain names to renew. If we think about the question in a quantitative fashion, the product of the annual probability of sale for the domain name times the expected net return from a sale of that domain name should be higher than the annual costs to hold the domain name in order for it to make sense to renew.

EXAMPLE â€“ Should I Renew This Name?
Letâ€™s say the probability of sale of a particular domain name at 0.25%, a bit less than the industry average. If we estimate the net return from a sale, if it happens, based on pricing, expected commission, acquisition cost, etc., would be \$1500.​
0.0025 x \$1500 = \$3.75​
Therefore it would not make sense to renew this particular domain name, assuming that the renewal fee was \$9.00.​

Note that the net return is your gross sales price minus commissions and other costs, and minus your acquisition cost.

I have not included parking revenue in the calculation. If your domain name earns enough parking revenue to cover renewal costs it of course justifies keeping.

If you have a large portfolio, and have been successful in domaining for some years, you can estimate the values for a domain name with more confidence than if you are just starting out.

I covered some factors to consider when deciding to renew a domain name in How To Decide What Domain Names To Renew.

Renew In Advance?

Verisign is raising the wholesale price on `.com` by another 7% at the end of this month. Some investors are renewing domain names in advance of the price increase. Others argue that if the name sells that is essentially lost money, and only renew near expiration.

Letâ€™s quantitatively look at the situation using the annual probability of sale of the name, p. If the name does not sell before the period when the renewal kicks in, you have saved the difference between the current and increased renewal rate. On average, the saving is p times the cost savings. However, there is a chance that the name will sell quickly, and you have wasted the amount spent. In this case, we multiply the current renewal times (1-p) where p has been expressed on the usual 0 to 1 scale. I summarize the two calculations below.

EXAMPLE â€“ Renew In Advance Of Price Increase
You hold a `.com` name that you have decided you will want to keep for more than one year if it does not sell. To get a representative cost, I took the 5th best current `.com` renewal rate at TLD-list, which turned out to be \$9.13. With a 7% increase, we expect the new retail renewal to be about \$9.77.​
Letâ€™s say it is a good name, and you think the probability of sale in any one year is 10%, much better than the industry average. So there is 1 chance in 10 the name sells during the next year before you need that renewal you did in advance. If we multiply \$9.13 x 0.10 = \$0.91 that would be the probabilistic loss.​
Now letâ€™s look at what probability suggests we save by renewing early. There is a probability of 0.90 that we will need to renew the domain name because it did not sell. We save \$0.64 by renewing in advance, so that would suggest a \$0.64 x 0.90 = \$0.58.​
Since the saving, for these numbers, is less than the \$0.91 calculated earlier, for this high probability of sale domain name it does not make sense to renew in advance.​
However, if instead we look at a more typical annual probability of sale, say 1%, the results are different. In this case \$9.13 x 0.01 = \$0.09 is the average loss by renewing names like this in advance. The expected average saving is now \$0.64 x 0.99 = \$0.64. Renewing in advance is a no-brainer in this case, if we have the funds and we are sure we want to keep it long term.​

Note this line of thinking does not apply to domain names that you do not feel confident are high enough quality to hold long-term.

Also, even when the calculation shows we save money, on average, by renewing in advance of price increase, only you can say if you have the cashflow to act on it. Even if you do have the funds, you need to ask if this is the best use of money you have. The funds you put into advance renewals will not be available for increasing your portfolio size through acquisitions, which might yield a better return.

Final Thoughts

There are lots of other examples of ways to apply probability ideas to domain name investing decisions.

For example, does it make sense to hold domain names with high premium renewal rates?

What about transferring a name to save on renewal cost, which may make the name not open to fast transfer networks for a period of time, depending on registrar. Do the savings justify the period when your sales chance will be decreased?

An interesting case is for `.co` and some new extensions where you can get the first year at a discounted rate compared to the regular renewal. The probability argument for this case may suggest the name is worthwhile to hold for one year, but not long term at the higher annual cost.

Perhaps the most important case is using probability estimates to decide whether to accept an offer, or hold out for a higher price, but with some chance the name will never sell.

What about a name you have decided to drop? Does probability suggest it is better to keep the domain name listed at retail price right to the end, or to liquidate for some return?

I may take up some of these other applications in a future article.

I hope you will share interesting ways you use quantitative thinking in your domain name decisions.

•••
The views expressed on this page by users and staff are their own, not those of NamePros.
My skill is street smart ...keeps me alive and smart

•••
Thanks Mr. @Bob Hawkes. It's always nice reading your articles.

•••
Great stuff as always Bob.

I frequently tell new people that they have to learn the numbers to have the best shot at a profit. STR, acquisition cost, renewal, and sale price (minus commissions) all have to align to see a profit.

I've taken a sobering look lately at things like domain length. When I ran the numbers from over 20 years of sales, I realized I was carrying some dead weight. The percentages just didn't work.

•••
Very interesting and incredibly detailed. Thanks a lot Bob!

•••
STR will vary greatly based on if a domainer sells wholesale or retail.

•••
Great article.

As I have said many times, as a domain investor you really need to play the likelihoods to build a repeatable business model.

It is hard to build a business on outliers.

Brad

•••
STR will vary greatly based on if a domainer sells wholesale or retail.
It will also vary greatly based on the quality of domain, exposure, price, etc.

You can still predict a relatively standard STR based on similar quality domains, prices, and exposure.

1%-2% seems like a pretty standard range when it comes to passive end user sales at normal pricing (compared to quality).

Of course this is more predictable with a larger sample size. It will likely be a lot more accurate with 10K domains than 100 domains, especially over time.

Brad

Last edited:
•••
Great analysis, as always, Bob.

Couple of comments:

- I largely agree with the estimate of around 15 MM .com for sale, +/- 10%, as Afternic has around 21-22 MM listed and I assume about 2/3 of those would be .com. Of course, not all names are on Afternic, but also not all names there are listed at \$1k+. So, maybe, that balances out.

- I don't think it is a good estimate about 25% of sales ending up on NameBio. I have sold around 200 .coms at above \$1k in 2021-2022 (19 months) and I don't think any of those sales made it to NB. I'd assume only around 10%-15% of sales make it to NB.

- If we do assume range of 15MM-18MM of names for sale and NB disclosed % at 10%-15%, then the approximate STR range would be in 0.55%-1%. Perhaps, closer to the lower value, due to the number of non-investment grade names among the 15-18 MM listed names. My "gut feel" estimate for the across the board STR is in 0.6%-0.75% range for \$1K+ listed .coms.

- STR should be calculated not by dividing by the average number of the domains in the portfolio, but by the weighted average, which could be different if your portfolio growth is not distributed evenly across the year.

•••
It is also mind-bogging to think that almost 50% of all listed .com names are owned by one company: HugeDomains.

Huge indeed!

I humbly own only about 0.15% of those.

•••
If we do assume range of 15MM-18MM of names for sale and NB disclosed % at 10%-15%, then the approximate STR range would be in 0.55%-1%. Perhaps, closer to the lower value, due to the number of non-investment grade names among the 15-18 MM listed names. My "gut feel" estimate for the across the board STR is in 0.6%-0.75% range for \$1K+ listed .coms.
Thanks as always for your well expressed and reasoned comments @Recons.Com I agree it is totally possible the fraction of \$1000 plus sales on NameBio is lower, which as you say would make STR higher. In past I have generally assumed 20%. I sense more investors are individually reporting sales to NameBio than used to be the case.

I perhaps should have stressed in article that my main focus was not to produce specific values, but rather to stress how easy it is for anyone to do their own analytics. I wanted to be transparent with the factors I used, but it is easy for anyone to redo with different factors, or look at different TLDs, price ranges, etc.
STR should be calculated not by dividing by the average number of the domains in the portfolio, but by the weighted average, which could be different if your portfolio growth is not distributed evenly across the year.
That was exactly what I meant by average. For example take the number of domain names you had each month, and take the average of the 12 to get approximately the right number for the year. Sorry this was not explicitly clear.

Thanks again for great points. Congratulations on all your sales.

Bob

•••
almost 50% of all listed .com names are owned by one company: HugeDomains
What number of names does HD have listed? According to their website it says 4 million. Is there a more precise number somewhere?

Whatever the precise number, it is sobering that a few companies account for so many of the aftermarket names.

Bob

•••
Excellent write up on sell through rates. New domain investors will benefit greatly from understanding this information, as many are disappointed at how few sales actually happen in a regular portfolio.

The key is quality over quantity in my opinion, if you want to succeed long term.

Last edited:
•••
What number of names does HD have listed? According to their website it says 4 million. Is there a more precise number somewhere?

Whatever the precise number, it is sobering that a few companies account for so many of the aftermarket names.

Bob

I should have been more precise. They probably own 50%+ of listed .com priced 1k+. They used to have around 7 million few years back. I assume they are at around 8 million now.

Last edited:
•••
Thanks as always for your well expressed and reasoned comments @Recons.Com I agree it is totally possible the fraction of \$1000 plus sales on NameBio is lower, which as you say would make STR higher. In past I have generally assumed 20%. I sense more investors are individually reporting sales to NameBio than used to be the case.

I perhaps should have stressed in article that my main focus was not to produce specific values, but rather to stress how easy it is for anyone to do their own analytics. I wanted to be transparent with the factors I used, but it is easy for anyone to redo with different factors, or look at different TLDs, price ranges, etc.

That was exactly what I meant by average. For example take the number of domain names you had each month, and take the average of the 12 to get approximately the right number for the year. Sorry this was not explicitly clear.

Thanks again for great points. Congratulations on all your sales.

Bob

Thank you for your compliment. It means a lot!

Weighted here is important. Just average might not do the trick in certain scenarios although monthly average might be good enough approximation.

•••
•••
Well then here's a question. What is the applied probability of the average domainer earning more profit from: Sales or development? Seeking calculations not opinions.

•••
Well then here's a question. What is the applied probability of the average domainer earning more profit from: Sales or development? Seeking calculations not opinions.
It is impossible to give a calculation when the actual numbers, and formula itself are unknown.

Domains have more of a fixed price than development. It is an impossible question to answer.

How long is a piece of string? There is not enough information there.

Brad

Last edited:
•••
Wow @Bob Hawkes, very in-depth. Enjoy your weekend

•••
It is impossible to give a calculation when the actual numbers, and formula itself are unknown.

Domains have more of a fixed price than development. It is an impossible question to answer.

How long is a piece of string? There is not enough information there.

Brad
An average for potential domain sales is the only thing which could be calculated, aside from pre-agreed domain sales. Conversely the calculation for domain name development is not impossible to answer, in fact with a consistent flow of traffic you can fairly accurately predict your income. What do you think?

•••
An average for potential domain sales is the only thing which could be calculated, aside from pre-agreed domain sales. Conversely the calculation for domain name development is not impossible to answer, in fact with a consistent flow of traffic you can fairly accurately predict your income. What do you think?
A domain purchase has a fixed cost.

Web development has an unknown cost. If can go anywhere from \$0 to hundreds of millions.
The profitability depends on the content, niche, search rankings, traffic, and many other factors.

Plenty of development projects fail. Many not only lose capital but also sweat equity (aka man hours) as well.

Circling back to your original question -

"What is the applied probability of the average domainer earning more profit from: Sales or development?"

Again, impossible to answer without more information.

Brad

Last edited:
•••
An average for potential domain sales is the only thing which could be calculated, aside from pre-agreed domain sales. Conversely the calculation for domain name development is not impossible to answer, in fact with a consistent flow of traffic you can fairly accurately predict your income. What do you think?

For .com investment (to keep it simple and not add the complexity of different and, even worse, varying renewals), all you need to know is:

- your investment amount (cost of acquisition)
- revenue (portfolio size * STR * average price post commission)
- renewals (total names * 9.xx)
- other costs, if any (subscriptions, overhead)

For the development though you need way more details and it could vary from project to project

- cost of development
- cost of acquisition
- cost of maintenance
- cost of updates (this and all above could vary drastically for each project)
- revenue from ads (need to make lots of assumptions. How many visitors per month? Does it change month to month? How to monetize? Which countries are visitors from? A visitor from US could be worth 50 times more than a visitor from a poor country earningwise)

In short, it is not apples to apples and, frankly, there is no point in such a comparison.

You might as well compare starting plumbing business with domain investment.

•••
For .com investment (to keep it simple and not add the complexity of different and, even worse, varying renewals), all you need to know is:

- your investment amount (cost of acquisition)
- revenue (portfolio size * STR * average price post commission)
- renewals (total names * 9.xx)
- other costs, if any (subscriptions, overhead)

For the development though you need way more details and it could vary from project to project

- cost of development
- cost of acquisition
- cost of maintenance
- cost of updates (this and all above could vary drastically for each project)
- revenue from ads (need to make lots of assumptions. How many visitors per month? Does it change month to month? How to monetize? Which countries are visitors from? A visitor from US could be worth 50 times more than a visitor from a poor country earningwise)

In short, it is not apples to apples and, frankly, there is no point in such a comparison.

You might as well compare starting plumbing business with domain investment.
For the average domainer, the odds of getting a return are FAR higher from parking or developing a domain name than selling a domain name. That's simple mathematics. The average person who tries to make a profit from selling domain names WILL fail, however hard they try. The average person who parks a domain name is highly likely to make at least some revenue, and probably profit over 12 months if they really research the name before registering. The average person who develops a 15 dollar H.R. domain name properly will DEFINITELY make profit over 12 months. Just need to earn average 30 cents per day to earn 109.50 dollars per year. The reason selling is pushed so much on Namepros and parking and developing isn't is simple - because you guys ONLY make money from selling your domain names. I get it know. I can see why there is so much hate, obstruction and denial, but it only serves to compound the misery of newbie's who don't realise that the system is rigged to make them fail. Some anonymous person will now downvote this post 10 points.

•••
Web development has an unknown cost.
Sorry but that's simply untrue mate.

•••
Plenty of development projects fail.
A properly developed 10 dollar domain name that doesn't earn 100 each year? It's 25 cents per day. Come on it's not rocket science. Set up takes what two hours. Ten or twenty minutes updating each day if that. I'm convince domainers are just addicted to the rush of a sell, the gambler's mindset. It's not really about creating a sustainable business model which can be repeated.

•••

4 comments
3K views
33 comments
3K views
11 comments
3K views
5 comments
2K views
10 comments
2K views