Selling Before Renewal: Turning Dead Weight Into Cash Flow

Our 177 priced names carry a median ask of $488 and a renewal bill near $2,124 a year at the cheap end. The arithmetic of a name that has to sell once every fifteen years to pay for itself, the four exits we run 90 days before the invoice, and the portfolio numbers behind them.

Every year the renewal invoice arrives as a single number, and every year most portfolio owners pay it without reading it. That number is the only place where the whole shelf gets priced against itself. It does not care what we believe about a name. It only asks whether we want to keep paying for it.

We treat renewal as a cash-flow decision rather than a housekeeping task. The reason is arithmetic: on our own shelf, the entire annual renewal bill is a rounding error against the asking prices, which is exactly why it hides. The problem is not the total. It is the bottom of the distribution, where names cost months of renewal to sell and years of renewal to keep.

What dead weight actually looks like in our own numbers

The public portfolio page at dnwai.com/d/ lists 178 names. Of those, 177 carry a buy-now price: 146 .com, 23 .si and 8 .net. The median ask is $488. The mean is $689. The whole priced shelf adds up to $121,872.

Now the distribution, because the median hides it:

36 names are priced under $200, which is 20% of the shelf.
13 of those sit at exactly $195, 9 at $150 and 8 at $100.
97 sit between $200 and $499, with 36 at $495, 28 at $375 and 20 at $488.
4 sit between $500 and $999, 34 between $1,000 and $1,999, and 6 between $2,000 and $4,999.

At the cheap end of renewal — call it $12 a name a year, which is what the least expensive extensions cost us, not a portfolio average — the shelf costs about $2,124 a year to hold. That is 1.7% of the asking total. Framed that way, renewal sounds trivial. Framed against the bottom band, it is not: the 36 names under $200 are where inquiry flow is thinnest and where the renewal is the largest share of the expected proceeds.

The only test that decides anything

A name is worth keeping when its expected net proceeds this year exceed its renewal. That gives a single ratio:

A $195 sale nets $180.38 after Atom's 7.5% portfolio commission. At $12 a renewal, that is 15 renewal years.
A $100 sale nets $92.50. That is 7.7 renewal years.
A $50 wholesale sale nets $47.50 at the 5% wholesale rate, which is just under four renewal years.

Invert it and you get a hurdle: a name priced at $195 has to sell more often than once every 15 years to pay for itself. That sounds like a bar so low it is meaningless, and for a name with a real buyer pool it is. But probability is not uniform across a shelf. For a name with no comparable sale in its extension, no inquiry in twelve months and no identifiable operator who would want it, the probability is not 6.6% a year. It is close to zero — while the renewal is certain.

That asymmetry is the whole argument. The renewal is a guaranteed cost against an uncertain return, and the only lever we control is whether we keep paying it.

The invoice is the wrong deadline

Deciding at renewal is deciding late, for three reasons.

The first is market time. A listing needs weeks to be seen, an auction needs a bidder who is awake that week, and a transfer needs a payment window. Thirty days before expiry there is no room for any of it. We start the decision 60 to 90 days out, which leaves a full auction cycle plus escrow.

The second is leverage. A buyer who can see that a name expires in three weeks knows the seller is cornered. That is not a negotiating position; it is a discount request waiting to be written.

The third is honesty. Renewal is the moment the sunk-cost reflex is strongest: we have held the name for two years, so paying once more feels like protecting an investment rather than starting a new one. The invoice makes this easy to fix by asking one question per name — would I buy this name today at its renewal price? If the answer is no, the price we paid is irrelevant.

Four exits, in order of what they return

We run the same ladder for every name that fails the test.

Outbound to the operator already using the name. Best price by a wide margin, highest cost in hours. This is the exit that turns a $195 ask into a real conversation, and it is the one that produced our only genuine human replies this year.

A self-set discount on the retail listing. Not a panic cut — a scheduled one. If a name has sat at $495 for twelve months with no inquiry, the honest read is that $495 is not the price; the sale we want is the one that clears before the invoice. Moving to $375 costs $120 of theoretical proceeds and buys a real chance of an exit.

Wholesale. Atom charges 5% on wholesale against 7.5% on retail. On a $195 name that is $185.25 against $180.38 — the wholesale route costs $4.87 and closes in days instead of quarters. For a name we have already decided not to keep, that trade is close to free.

A $50 no-reserve auction. It always sells, which is the point, and it can clear at $10, which is the risk. A $10 result nets $9.25 — less than one renewal. Liquidation is not a recovery; it is a stop on a recurring charge, and it should be judged that way.

And the fifth option, which nobody lists because it feels like failure: let it expire. When the probability really is near zero, expiry is the cheapest exit available. It costs nothing, returns nothing, and stops the bill.

The fixes that make it worse

Raising the price to justify renewing. A higher ask lowers the chance of a sale faster than it raises the net. A name at $495 with no traffic does not become viable at $695.

One more year. Each year is a fresh $12 and a fresh 6.65% hurdle. Two extra years on a $195 name is $24 against a $180 net — the name now has to sell once in thirteen years just to break even on the delay.

Renewing by autopay because the invoice is one line. This is the real mechanism behind dead weight. One payment method, one renewal date, one number: nothing in that flow ever asks a name to justify itself.

The calendar as a budget

What changed for us was not a pricing insight, it was a process. Renewals are grouped by month and reviewed as a batch, 90 days ahead, with three fields per name: last inquiry date, best comparable sale, and whether a named buyer exists. Names under $200 with none of the three go to liquidation in that window rather than to another renewal.

The cash-flow logic is simple. A $195 name sold before renewal pays for fifteen renewals of a $12 name — the same $195 held as inventory pays for none. Turnover is not a defeat; a shelf that never loses a name is a shelf that never tests its prices.

We also track three numbers now, because they are the ones that move: retention (share of names renewed), exit mix (how many left by sale versus expiry), and the liquidation share of total exits. A portfolio where liquidation is 30% of exits and retention is 95% is telling you something the asking prices are not.

What we still renew without argument

The ladder is not a rule to apply blindly. We keep, and keep paying for:

Brandables with a defined buyer pool, even at a $1,795 ask, because the buyer arrives on their own schedule.
Geo and category names where a local operator is already trading under a near-match.
.si names with a matched operator, where the extension is the whole story and the buyer is one email away.
Anything with a live thread, an open offer, or a pending appraisal we have not answered.

The judgement is not about optimism. It is about whether we can name the buyer. If we can, the renewal is an investment. If we cannot, it is a subscription to a hope.

A seven-day pre-renewal routine

Pull every name renewing in the next 90 days, with price and last inquiry date.
Mark the ones with no inquiry, no comparable and no named buyer.
Sort into keep, discount, wholesale, liquidate.
Send the liquidate list to a $50 no-reserve auction with a seven-day clock.
Drop the discount list by one band, once, on a schedule.
Write the outcome down. A decision that is not recorded gets relitigated next year.
Put part of the proceeds into one better name instead of renewing three weak ones.

None of this requires a forecast. It requires a date, a price and a yes-or-no answer per name. The invoice will arrive either way; the only choice is whether it arrives as a surprise or as a budget line we already decided.