Sizing your first order without guesswork

Size your first order from the sites you can name today, not the portfolio you hope to own in a year. Count what is live, group those sites by whether they may be linked back to each other, and buy one address per group. The result is nearly always smaller than a first-time buyer assumes.

Step 1: build the inventory before you look at any plan

Open a blank sheet and give every domain you control one row. For each row, record three things:

1. State. Live with real pages, half-built, or parked with nothing on it.
2. Owner or client. Who the site belongs to commercially, which is not always you.
3. Exposure. Whether the site openly carries your branding, or is meant to stand on its own.

Then apply one filter: a row counts toward this order only if it is live now, or the content is ready and a launch date sits inside roughly sixty days. Parked domains and "maybe next quarter" ideas are wishes, not capacity requirements.

Step 2: split the counted rows into isolation groups

Two sites belong in the same group when it would cost you nothing if somebody noticed they sit on the same address. They belong in different groups when that discovery would be a problem. Work through your rows against this:

Be honest in the third column. A separate address is a footprint control and a blast-radius limit. It is not a ranking input, and no count of them will lift a page on its own.

Step 3: a worked example

Suppose your sheet holds 21 domains. Nine are parked or unwritten, so they drop out and 12 remain. Of those 12:

3 client sites, 3 separate businesses    -> 3 groups
1 agency site + blog + docs subdomain    -> 1 group
6 niche sites, one visible brand, openly cross-linked -> 1 group
2 review sites meant to read as independent -> 2 groups
                                            ---------
                                              7 groups

So the order is seven addresses, not 21 and not 12. If two more client contracts are already signed with launch dates in the next two months, add those two and order nine. Do not add a buffer for contracts you are still pitching.

Step 4: pick the plan the group count implies

Groups tell you how many addresses. Traffic and software tell you how much machine. A dozen brochure sites on separate addresses are lighter than one busy store, so decide the two independently. Entry-level multi-IP hosting covers most first orders; move to a VPS only when a real resource ceiling, not the address count, is what pushes you there. How those addresses should be spread across ranges is a separate decision, covered under network planning.

Step 5: sanity check before you pay

• Every address on the order maps to a named group on your sheet, with no spares "just in case".
• Every counted site has content that exists right now or a dated launch plan.
• For each split you can state what would go wrong if those two groups shared an address. If you cannot, the split is imaginary and they merge.
• The plan came from load and the address count came from grouping. If one number drove the other, redo the split.

Under-ordering is cheap to fix, since addresses can be added later without moving anything. Over-ordering is a recurring charge for an estate that never arrived. Order the number you can defend line by line, then revisit the sheet when a real launch lands.