
Atom Premium, priced honestly: the commission ladder, the NS rule and 13 listings with real numbers
Thirteen domains in my portfolio went from wholesale prices to retail ones this month, and the decision that moved them was not the price. It was the marketplace programme I put them into: Atom Premium. Premium comes with a rulebook — a name-server requirement, a commission ladder, an exclusivity clause and a notice period — and that rulebook decides what you keep from every sale long before a buyer shows up.
So this is not a review of the badge. It is the arithmetic: what the ladder pays at each band, what the name-server requirement costs you, where the AI appraisal belongs in the decision, and the two mistakes that quietly cancel a premium listing — one of which I made myself this week.
What Premium actually is
Premium is a curated retail programme, not a listing tier. You submit a name and Atom approves or rejects it; approved names are marketed with paid campaigns, retargeting, priority placement in marketplace search, syndication to partner marketplaces, a brokerage team that negotiates with buyers, and a branded landing page built on your behalf. Instant Premium is the faster lane — no AI credit needed, instant approval — with the same obligations once you are in.
The obligations matter more than the marketing:
- Your name servers must stay on Atom at all times while the name is listed. The landing page, the tracking and the verification ride on that delegation. Move the NS and the listing comes down — and repeated removals can limit future submissions.
- Removing a name needs 30 days' notice, so it is not a programme you dip in and out of over a weekend.
- Exclusivity: a Premium listing is meant to live on Atom. The partial-exclusivity policy allows a handful of approved partner marketplaces, and Atom may mark your name up there — it does not allow you to undercut your own Premium price somewhere else.
- Minimum retail transaction $100, which tells you what kind of inventory the programme is built for.
The commission ladder, in real money
The published Premium ladder for a standard listing runs on bands:
- 30% on everything up to $4,998
- 25% from $4,999 to $49,999
- 20% from $50,000 to $75,000
- 15% above $75,000
Long-term listings earn a lower rate: once a name has been continuously listed for two years, the sub-$2,499 band drops to 25% with an AI-credit submission (30% on the free route). If you submitted without a credit and without an instant approval, read your own dashboard before you price anything — the legacy rate on that route has been as high as 35%, and one percentage point at these numbers is not small. Confirm the rate you are actually on, then do the arithmetic with that number.
Which is the point: the list price is not the number that matters. The net is.
- List $1,195 → $836 in your pocket at 30%
- List $1,795 → $1,256
- List $2,095 → $1,466
- List $2,495 → $1,746
A $195 buy-now that actually sells pays you $136. A $2,095 listing that takes four months pays $1,466. The programme is not built for small numbers — it is built for names you are willing to hold at a retail price.
The appraisal is a ceiling, not a target
Instant Premium arrives with an AI appraisal, and the tempting move is to copy that number into the buy-now field. I use it the other way round. The appraisal is the ceiling of what the marketplace believes the name is worth — and Atom prints it on your own landing page as "Estimated value", right next to your asking price. Set your price ten per cent under that figure and the page reads as a bargain before the buyer has read a word of your description; set it above and you are asking someone to pay over the marketplace's own estimate.
Here is what that looked like across thirteen names — the old price, the appraisal Atom showed, and the BIN I set:
- Solarell.com — was $195 · appraisal $2,399 · now $2,095
- AlwaysTherapy.com — was $375 · appraisal $2,099 · now $1,795
- LemonWhite.com — was $100 · appraisal $2,799 · now $2,495
- LocalSafes.com — was $375 · appraisal $2,299 · now $1,995
- Flyprinting.com — was $150 · appraisal $1,999 · now $1,795
- TriRack.com — was $150 · appraisal $1,899 · now $1,695
- Pharmacco.com — was $150 · appraisal $1,999 · now $1,795
- TrendyTreatment.com — was $195 · appraisal $999 · now $895
- SkinCareTraining.com — was $375 · appraisal $1,399 · now $1,195
- ParalegalLab.com — was $2,195 · appraisal $1,499 · now $1,295
Two of those went down, and that is the point
ParalegalLab was priced above its own appraisal, and TrendyTreatment was sitting at $195 while the marketplace valued it just under a thousand — a price that tells retail buyers the name is junk. Repricing is not a ratchet upwards. It is aligning the ask with the band the marketplace is willing to argue for, and being willing to cut a name that sits above it.
For the weaker names I deliberately stayed under the appraisal rather than at it. A listing that reads as a bargain converts faster than one that reads as fairly priced, and at these bands the difference between $895 and $1,000 is thirty dollars — not worth a longer hold.
The mistake that cancels the premium: two prices for one name
My own trap was not the pricing. It was that the same thirteen names were still listed at $100–$495 on a wholesale liquidation thread and in my bulk spreadsheet while Atom was asking $1,795 for one of them. That is a cannibalised listing: any trade buyer — and a fair share of retail buyers — takes the cheap one, and the retail price becomes decoration. Worse, a buyer who finds both learns your floor and negotiates down to it.
The rule I hold to now: one name, one channel, one price. A name in a retail programme leaves the liquidation lane entirely. If you want to liquidate, liquidate — with names that are not on the retail shelf.
What the name-server switch really costs
Pointing a domain at a marketplace's name servers is never just a DNS change. Two things break, and both broke for me:
- Email on the domain stops. If the name carries a mailbox or a registrar forwarding rule — MX records pointing at eforward1-5.registrar-servers.com, or mail hosted on your own server — that mail dies the moment the delegation moves. I checked before switching and still found two names with live forwarding; the move had to be a knowing decision, not a surprise.
- Verification tokens disappear. TXT records other marketplaces used to verify you — Afternic's afternic-verification-… token, for one — live in the same zone and vanish with the delegation, so a verification can lapse. Snapshot every record you care about before touching the NS, and restoring one becomes a single command instead of an archaeology exercise.
The cheap version of that lesson: dump NS, MX, TXT, A and CNAME for the whole list into one file first. It takes a minute and it is the difference between a two-minute rollback and a lost mailbox.
Pre-flight, before you list anything
This week I nearly published a warning that one of my own names had dropped — because I checked a spelling that was one letter off. The name in my portfolio is predictivemortgage.com; the string I tested was predictivmortgage.com. Two different names, two different answers: the first resolves through Atom's name servers with an expiry in 2027, the second does not exist at all. A listing built on the wrong string sends buyers to a dead page.
So every name now gets four checks before it enters any programme:
- Is it registered? Ask the registry, not your memory: an RDAP lookup returns 200 for a registered .com and 404 when the name is not in the registry at all, and an NS query returning NXDOMAIN tells the same story.
- Is it yours? Read your registrar's portfolio list — not a spreadsheet row, a folder name or a landing-page path. Ownership lives with the registrar.
- Is the spelling exactly right? Compare the owned name with the name in your campaign copy as strings, character by character. One letter caught earlier is far cheaper than one letter caught later.
- When does it expire? A retail listing is a promise that you still hold the name in four months. Six of mine renew between November and December; I would rather know that today than after a sale I cannot complete.
Does Premium sell faster? An honest answer
Nobody publishes a sell-through rate, and the silence is the answer. What Premium buys is audience, retargeting and a brokerage team — not a buyer. In my own dashboard a dozen freshly approved listings collected a few hundred views between them in their first weeks and zero shortlists, which is where most early listings sit.
What actually moves a name inside the programme:
- a price inside the band the market will argue about, as above;
- negotiation switched on, with a floor you decided in advance instead of in the moment;
- auto-pilot discounts and exit-intent offers enabled — the marketplace notifies people who already looked, and a large share of retail sales come from a second look;
- and a landing page that answers the buyer's question — the name, the price, the transfer, the escrow — in one screen.
The expensive version of this is expecting the programme to do the pricing for you. It will not. It markets the number you choose.
The checklist I now run on every name
- Registry check (RDAP 200), ownership check in the registrar wallet, spelling check — and the expiry date written down.
- DNS snapshot: NS, MX, TXT, A, CNAME, before any name-server move.
- Choose the channel: retail programme or liquidation lane, never both for one name.
- Set the BIN about ten per cent under the appraisal, and write down the floor you will accept before the first offer arrives.
- Make that the only number: portfolio page, spreadsheet, outreach copy and every marketplace listing.
- Re-check the live listing after publication — a price is not set until someone can see it.
None of this makes a name sell. It makes the moment it sells worth having — and it stops a marketplace's rules from quietly eating the price you spent months justifying.
Want a shortlist instead of a lecture?
160 buy-now priced domains — brandable, exact-match and geo — with escrow and same-day transfer. Tell me the industry and I will send the three names that fit: ask for a shortlist.