Ads stop when you stop paying. A good page keeps earning.

What it does

1. Who writes it?

   A second pass, separate from the one that wrote the post, aimed at 140 to 155 characters. Long enough to say something, short enough to survive the display limit.

2. What if that pass comes back too long?

   It is cut to fit before it reaches your store, so the trimming happens here on a finished sentence rather than in a live listing.

3. What if it fails outright?

   The post's opening paragraph is used instead, trimmed to the same limit. Nothing publishes with the field blank.

What you end up with

1. Where does it actually end up?

   On the article's description field in your store, which is what your theme hands to search engines. Set at publish, not left on a list for you. YOUR STORE

2. Does it describe the post that exists, or the one that was planned?

   It is written from the finished draft rather than from the brief, so it can only describe what actually got written.

What it does

1. What decides it?

   The topic the post was planned against, so the page is titled with the subject it was commissioned for.

What you end up with

1. Where does it end up?

   On the article in your store, which is what your theme prints at the top of the page. YOUR STORE

What it does

1. Where does the address come from?

   The post's own title, reduced to lower-case words and hyphens, so what a reader sees in the bar is the subject rather than an identifier.

2. What if that address is already taken?

   Your live store is checked before anything is created. An article we did not write is never overwritten. YOUR STORE

3. What if the check itself cannot complete?

   Then nothing publishes. An address we could not verify is treated as unsafe rather than assumed free. YOUR STORE

What you end up with

1. So what happens to the post that collides?

   It goes live at its own address with a short suffix worked out from the post itself. The same post always lands on the same URL, so a retry cannot create a second copy.

2. Does editing a published post move it?

   No. Updates deliberately never resend the address, because a URL that moves gives up whatever it had earned.

What it does

1. What stops a run spending without limit?

   A fixed ceiling per run, checked before each post is written, against the token counts the model actually reports rather than an estimate.

2. How many rewrites do I get?

   One full AI revision per post, held on the server rather than in the interface. Editing the words yourself is unlimited and does not spend it.

What you end up with

1. Which instructions wrote this particular post?

   Stamped onto every draft. When the writing changes for the worse, the version that wrote it is on the record instead of being guessed at.

2. What has to be true before it can publish?

   Six hundred words counted from the finished HTML, a title, at least one paragraph, and no leftover placeholders. A draft that misses any of those is refused.

3. Can it publish without me?

   No. Writing a post touches your store zero times. Publishing is a separate action you take, and it creates a draft in Shopify unless you say otherwise. YOUR STORE

What it does

1. What is a pillar here?

   A named subject with its own clusters, stored for the store and reused across campaigns rather than rebuilt for each one.

2. What does it do to a plan?

   Planned topics are filed against a pillar, and each pillar keeps a count of how many campaigns have used it.

What you end up with

1. What do you end up holding?

   A standing plan you can add to and reuse, with the counts on it, instead of a topic list that starts from nothing every time.

What it does

1. Where does the voice come from?

   A tone profile you set once and keep in Settings, not a phrase retyped into each request.

2. When is it applied?

   It is written into the instructions before the draft is generated, not edited in afterwards.

What you end up with

1. What changes when you change it?

   The next post, rather than a pass back over the last one.

What it does

1. What is compared?

   Every planned topic is compared against the posts already published for you, and against the rest of the same plan, before drafting starts. YOUR POSTS

2. How close is too close?

   Overlap is measured on the words that carry meaning in each keyword and the filler is ignored. Past sixty percent shared, the topic is flagged. YOUR KEYWORDS

3. When does it run?

   At planning time, before a word is drafted. Finding the overlap after both posts exist means one of them was paid for twice.

What you end up with

1. Does a flag stop the post?

   No. It flags and you decide. You can clear a whole plan of flagged topics in one action, or keep one and drop the other.

2. What is the check blind to?

   Pages we did not publish. It reads the posts written through PillarSEO, so anything that predates the connection sits outside it. YOUR POSTS

What it does

1. What runs, and when?

   A scheduled job pulls your Search Console rows once a week and stores a dated snapshot. SEARCH CONSOLE

2. What is actually being stored?

   Average position as Search Console reports it. That is not a live rank check, and the product does not have one. SEARCH CONSOLE

What you end up with

1. What is the history for?

   Giving this week something to be compared against, which is the only thing that makes a drop visible at all.

What it does

1. Where do the numbers come from?

   Straight off your Search Console rows, per page and per query, with no modelling in between. SEARCH CONSOLE

What you end up with

1. What can you see?

   Both numbers against the page they belong to, so a page that is seen and never chosen reads as exactly that.

What it does

1. How would you know?

   Your Search Console positions are pulled on a weekly schedule and kept as a running history, so this week has something to be compared against. SEARCH CONSOLE

2. What counts as a slip?

   Three positions or worse between snapshots. A one-place wobble is weather, and flagging it would bury the drops that matter. SEARCH CONSOLE

3. What about a page that held position and stopped earning?

   Also caught. A quarter of the clicks gone, or a click rate well under what that position usually returns, counts even when the number itself held. SEARCH CONSOLE

What you end up with

1. How long before it can tell me anything?

   Two snapshots, so roughly two weeks from connecting. Before that there is no history to compare against, and it says so rather than showing you a number. SEARCH CONSOLE

2. What does it do about it?

   It ranks what slipped and routes each one: refresh the writing, fix the links, check the technical side, or leave it alone. Starting the rewrite is yours. SEARCH CONSOLE

What it does

1. Which page brought the order in?

   Orders are matched to the page the visitor landed on, read from your store, and joined to that same page in your search data. YOUR ORDERS

2. How do those two ever line up?

   Addresses are put in one form first: casing, trailing slashes, tracking parameters. Skip that step and almost nothing matches, and the report reads as zero. SEARCH CONSOLE

What you end up with

1. How far back does it look?

   Thirty days, last touch, on the landing page. Someone who read a post in March and bought in June is not counted, and no model here pretends otherwise. YOUR ORDERS

2. What about revenue it cannot place?

   Search orders that do not match a known page are reported as their own figure rather than folded in. The gap is shown, not smoothed over. YOUR ORDERS

3. Does this decide what gets fixed first?

   No. The fix queue is ordered by how far a page fell and how much demand sits behind it, not by attributed revenue. The two are reported next to each other rather than blended into one number nobody can check.