How to buy a small online business

Browsing first is the single most expensive habit at this end of the market, and the reason is mechanical rather than psychological. What you see while browsing is asks. An ask is a number a seller chose; a comp is a number a buyer paid. Spend an evening reading asks and your sense of “normal” is now calibrated to the seller's side of every sale that has not happened yet — including the ones that never will. Every offer you make afterwards is an adjustment away from that anchor rather than a number you arrived at.

The alternative costs twenty minutes. Decide what you are shopping for — asset type, size of cheque, what you would do with it — and then look at what sales in that band have actually cleared before you look at a single listing. Our sold-price page is the dollar view of that, by asset type, read live from the same endpoint this page uses; asking price vs sold price is why the two numbers keep getting confused, including the four ways an ask reaches you dressed as a sale.

Loading live coverage from /api/stats…

The two comp columns count sales inside the window a report is built from — the most recent 200 for that type — that published enough for a multiple to exist on that basis. They are not the same as the rows column, and the gap between them is the whole story: most sold prices reach the public record with no earnings figure attached to divide by. Where a basis holds fewer than five such comps we publish nothing on it rather than a median a single new deal would move, which is the floor how many sold comps a valuation needs argues for.

Most listings you open are not deals, and the cheapest diligence is the kind that happens before you invest any. Three checks do most of the work, and all three are answerable from the listing page itself:

1. Is the headline number a sale or an ask? On a live listing it is an ask by definition — but the comparisons a listing quotes at you are frequently asks too, presented as market evidence. A number with no stated buyer and no stated close date is not a comp.
2. Does the earnings figure state its period and its basis? “$1,200/mo profit” with no window, no definition and no add-back list is not a figure you can price against; it is a claim you would have to reconstruct from scratch. Monthly vs annual multiples is where most quoted numbers go wrong, usually by a factor of twelve, and what counts as earnings is the rest of it.
3. Does the asset actually transfer? Traffic that depends on one account, revenue that runs through a rail in the seller's name, a brand whose usable part is a social handle — each of these can be genuine and still not arrive intact. What a headline sale price includes is the inventory to read a listing against.

The due diligence checklist for sub-$100k deals is the long version of this step, including the five-minute screen and what to ask for once a listing survives it.

Buyers new to this size spend their energy on the multiple, because that is the number that looks negotiable. It is the wrong end of the ratio. A multiple applied to an unverified earnings figure is a precise-looking answer to a question nobody has established, and the error in the denominator is routinely larger than anything you would win by arguing the multiplier down a quarter turn.

This is also visible in our own numbers, and it is worth being exact about what it does and does not show. Several asset types publish a median profit multiple below 1×: sold prices clustering in the hundreds against stated annual profits in the thousands. The honest reading is not that these businesses are worth two months of their earnings. It is that the market is pricing the claim — discounting an earnings figure nobody could check — and that a ratio can move because either end moved. What a profit multiple actually means takes that apart, and why similar businesses sell for different prices shows how wide the spread around any median really is.

So: get to the numbers behind the claim before you get to the price. Analytics you can log into rather than screenshots, payout records from the rail itself, a revenue history long enough to show a trend rather than a good month. Everything a seller cannot show you is a discount you are entitled to ask for, and the seller's side of this page spells out exactly what a well-prepared seller should already have ready — which makes it a useful list of what to request.

An offer that arrives as a single number invites a single counter-number, and the negotiation becomes a split-the-difference exercise with no reference to anything. An offer that arrives with the comp set behind it — this is the size band, these are the sales it is drawn from, here is where your deal sits inside it and why — is an argument the other side has to engage with on the evidence. It also protects you: writing the band down forces you to notice when your number came from enthusiasm rather than from comps.

How to make an offer is the mechanics, including the arithmetic behind the band and three sanity checks before you send it. Two of those checks are worth repeating here because they are the ones buyers skip: is the comp set your size (a set four times larger is describing someone else's deal), and is it recent enough to still describe the market you are buying in — how recent a sold comp needs to be puts a number on that.

One more test before you send: divide the price by the monthly earnings you actually believe, not the one in the listing. That is your payback period, and at this size it is a better instrument than the multiple because it is denominated in something you can feel — payback period on a small online business.

The close is where a good deal becomes a loss, and the failure mode is almost always ordering rather than fraud: funds moving before assets do, a handover “in principle”, a support period nobody wrote down. Four things to settle before anything moves, all of them cheap now and expensive later:

• Escrow, at any size. A refusal to use it is itself an answer, and the fee is small against what it prevents.
• A staged handover. Assets in your control, then funds released — and an explicit list of what “assets” means, account by account.
• The claims you are paying for, written as warranties. If the earnings figure is the reason for the price, it belongs in the agreement, not only in the listing.
• The support window, in hours and weeks. What the seller will do, for how long, and what happens if it does not arrive.

Settle in advance how money already billed before the handover is treated, and who carries cancellations and disputed charges that land after it. It is a five-minute conversation before close and a genuine dispute afterwards. The checklist's close section carries the full list, and what it costs to sell is the fee side of the same transaction, which is worth reading even as a buyer: it tells you what the seller nets, and therefore where their real floor sits.