Building a domain portfolio without losing money

Most domain portfolios lose money. Slowly, undramatically, in annual increments, through renewal fees on names no buyer has ever asked about. The failure is rarely one catastrophic purchase. It's four hundred mediocre ones, each perfectly defensible in isolation, all renewing on schedule against a sell-through rate that never turns up.

The discipline here is arithmetic long before it's taste. So start with the arithmetic.

One inequality decides the whole business

A portfolio makes money when annual sales revenue clears annual holding cost plus amortised acquisition cost. For a single year:

• Holding cost = number of names x average renewal fee
• Sales revenue = number of names x sell-through rate x average sale price

Cancel the count from both sides. What's left is the only ratio worth watching: sell-through rate x average sale price must exceed your average renewal fee. Size doesn't rescue you. It multiplies whichever side of that inequality you happen to be standing on.

Put plausible numbers to it. A .com renewal sits somewhere in the low teens of dollars once the ICANN fee is included. Hold 500 names and you're looking at roughly $6,000 to $7,500 a year in renewals alone. The commonly cited sell-through rate for an average portfolio is about 1% annually, which gives you five sales. For those five to cover the renewals before you've paid yourself a penny, they need to average $1,200 to $1,500. Then take off marketplace commission: on Names.com that's 15% of the completed sale, paid by the seller, so the gross has to be higher again.

Now nudge one variable. At 2% sell-through, you need half the average price. At 0.5%, double it. Tiny movements in sell-through swing the entire enterprise, and sell-through is decided almost entirely at the point of purchase, not by how energetically you market afterwards.

Every purchase needs a buyer attached to it

The cheapest available improvement to your sell-through rate is to stop buying names that will never sell. Before money moves, name the buyer. Not a demographic. An actual company, or an actual founder profile, with a specific reason to want this specific string.

Filters that reliably raise the hit rate:

• Does it describe a business someone is already running? Names built from live commercial vocabulary have a standing audience. Names built from a trend that peaked last year have nobody.
• Can it be said down a phone line without spelling it? Hyphens, doubled letters, digits and homophone traps each shrink the buyer pool.
• Will your buyer accept the extension? .com still carries the widest default trust. Alternatives can work when the word is exceptional or the extension is native to the sector, but assume a smaller, slower market.
• Would you use it yourself? If you wouldn't put it on your own company card, be honest about why anyone else would.

Names that clear all four are expensive. That's rather the point. Renewal cost scales with count while revenue scales with quality, which means two hundred good names beat two thousand weak ones on both sides of the equation.

Buying at one tier and pricing at another

This is where portfolios quietly break. There are broadly three tiers, and they behave nothing alike.

Exact-match commercial keywords

The generic word or phrase that is the product, of the sort covered in this piece on the keyword domain as a brand asset. Demand is real and recurring, because every new entrant in the category wants one. They cost the most up front and they move at reasonably predictable prices. Portfolio economics work best here.

Category killers

The short, dominant term for a whole market — the territory of category killer domain names. Low volume, long holds, occasionally enormous outcomes. These are not filler. One or two can anchor a portfolio. A portfolio made only of them will sit idle for years while the renewals stack up.

Brandables

Invented, pronounceable words with no inherent search demand. Cheap in, cheap to hold, and sell-through depends entirely on how good the word actually is. The brutal part: most invented names aren't distinctive enough to be worth anything to anyone, because a founder can invent a fresh one over a long lunch and register it at cost. Work this tier only if your edge is genuine linguistic quality rather than volume.

A price 30% too high is a price nobody answers

Every unsold name is paying rent. Overpricing doesn't delay a sale so much as prevent it, because the buyer never enquires at all. They see the number, conclude the project is unaffordable, and go and register something worse for $12.

What works:

• Publish a price. Names with visible buy-now pricing sell more often than names marked "make offer". Most buyers won't open a negotiation they can't size.
• Price against comparable sales. Your acquisition cost is sunk and the market has no interest in it.
• Offer instalments higher up the range. A founder who cannot sign off $40,000 can often approve monthly payments out of an operating budget. Structures like monthly payment plans on keyword domains rescue enquiries that would otherwise die on price alone.
• Use escrow, every time. Transfers on Names.com run through escrow, so neither side pays or hands over the name first. Deals collapse over trust more often than over money.

Forty-five days before the renewal batch

Set the reminder. Then go name by name and ask one question: if this weren't already mine, would I buy it today at its renewal price? If the answer is no, drop it.

Two rules make that easier to obey. A name that has attracted zero enquiries in three years is giving you factual information, not a temporary market mood. And sunk cost is not an argument: a name you paid $2,000 for and cannot sell is still costing you a renewal every year, and holding it doesn't recover a cent of the $2,000.

On a young portfolio, expect to cut 10% to 20% annually. That's the pruning that keeps average quality climbing.

Keep books, or you're guessing

Per name: acquisition cost, acquisition date, renewal date and fee, every enquiry with the offer attached, and the outcome. Without that ledger you cannot calculate your real sell-through rate, and without that number you're guessing at the single figure that determines whether any of this is worth doing.

Hold the commission and the tax position in view too. A $10,000 sale at 15% commission nets $8,500 before tax, and that $8,500 has to carry the renewals on every name that didn't sell that year. Portfolios are judged on annual net, not on headline sales.

What this business actually is

A low-liquidity asset business with a fixed annual carrying cost and a long tail of names that never move. It rewards patience, capital and a narrow buy filter. It punishes volume and optimism, usually in that order. If you can't state your target sell-through rate and average sale price today, don't scale anything. Buy five good names, hold them two years, and find out what your real numbers look like.