Internal Linker

I would not fix the timeout, I would delete the requirement. Ninety seconds of work inside a request means the user is staring at a spinner, any proxy or flaky connection between them and you can kill it, and a retry runs the whole thing again. The shape that has never let me down is: the action writes a job row and returns immediately, a background worker or a queue consumer does the render and the upload, and the client polls or subscribes for the result. It is an afternoon of work and it makes the whole class of problem go away, including the ones you have not hit yet.

The uncomfortable version of the rule: the reason the charming stories feel true is that they are better constructed than the truth. Real etymologies are frequently boring, frequently uncertain, and often end with scholars disagreeing. Anything that arrives complete, with a beginning, a middle, a punchline and no loose ends, has been shaped by retelling. Actual history has loose ends.

That explains why our litigation team kept quietly correcting the stop back out and never said why.

Turned toward getting to set against is the piece that makes it click. Facing is doing all the work.

Then go look at who's above you now and what the result layout looks like. A jump from 2 to 9 in nine days with no site change is usually either a competitor refresh or the SERP itself changing shape.

Start by confirming it's a ranking loss and not a demand loss, because they look identical in a sessions chart. Pull impressions and average position for that exact query in Search Console and compare year over year. If impressions held and clicks fell, you slipped in the ranking or lost a feature you used to hold. If impressions fell too, fewer people searched, and June is a real seasonal trough for a lot of topics. I've watched three people rewrite a perfectly good page because of a seasonal dip.

Timestamped sections in the description made a real difference for me here. People jump to their own problem and skip the intro, which is the part nobody wants anyway.

The pricing side is real too and it has a name. A more expensive item on the board makes everything next to it look reasonable, which is why the premium option often sells less than it appears to be designed to sell. Its job is partly to be the thing you decide against.

Bump products per page to 48 and add curated collection pages for the ways people actually shop. Depth 9 to depth 3 is a bigger lever than anything you can do to the product copy.

You've basically answered it in the last sentence, but the technical half people miss is internal linking depth. Pull a crawl and check how many clicks from the homepage each unindexed URL is. On most migrated shops the indexed ones are the ones linked from a category page in the main nav and the rest are only reachable through pagination or a sitemap, which is the weakest possible signal. Get every product within three clicks and linked from at least one hand-built page and a chunk of that 300 will start moving within a few weeks.

One thing that isn't time: sales. Ask whoever sells whether they send prospects a link to any of those 40 pages. I've consolidated away a page with 11 sessions a month that turned out to be the page the sales team pasted into every proposal. Traffic isn't the only reason a URL exists.

I bought one of those thousand fascinating word origins hardbacks as a present and then read it myself and it was full of the exact stories this room exists to debunk. The tell in the shop is the absence of dates. A book that gives you first attestation dates and says which language the word came through is doing etymology. A book that gives you an anecdote about a tavern in 1690 and no citation is doing entertainment, and there is nothing wrong with that except that people repeat it at dinner.

Flip-flopping is the classic signature, yes. Before merging, check the query data for each URL separately. If the 2024 post gets 80% of its clicks from a different set of long-tail queries, merging throws that away for a head term you might still not win. The test I use: if I removed one page entirely, would I lose any query I currently rank for? If no, merge. If yes, differentiate hard and stop competing.

Then retitle the 2024 one so the head term isn't in the title or H1, and link from it to the guide with the head term as anchor. You're picking a winner rather than hoping one emerges.

Twelve posts is also just not very many. A site with 12 pages has almost no internal linking to work with and nothing that looks like coverage of a topic. I wouldn't expect much before 30 or so posts in one tight subject area.