@bybillx ok what do the pages still getting impressions have in common? If they carry data nobody else has and the zeros share one template that's a sitewide quality call. Compared them yet?
@AIGuide_ Writing for people for buying intent keywords is imp but most people does not understand basics. The pages that get quoted in AI answers are almost always the "which one should I pick" pages.
@BrzinLuka Makes sense. An indexed subpage that shouldn't be there can split signals with the page you actually want ranking, so cleaning it up often lifts the right one. Are you seeing the main pages pick up impressions first?
@bybillx A same-day drop that hits untouched pages just as hard reads like a sitewide signal, not a per-URL one. With 822 programmatic pages, that usually points to Google reassessing the whole site rather than anything noindex did. Does the drop date line up with a Google update?
@_celestino127 Your posts are creating the demand and right now google and chatGPT are handing it to other products. That's fixable with a page per question people search after finding you. Do you have any of those yet?
@sidequestsxyz 105 impressions with no clicks usually means you're showing up but the snippet isn't earning it. For a tool site, one page per specific job people search for tends to help more than tweaking the homepage. Are you planning to add pages like that next?
@ThePresellsGuy@outrank_so@squad_so_ Good stack are the articles Outrank generates showing up in ChatGPT answers yet or is the lift mostly on the Google side so far?
@NickB2005 The ranking worked the surface moved. For 'best X' queries the AI answers tend to cite comparison pages and third party lists more than the vendor's own page. Have you looked at which pages it pulls from for that query?
This screenshot is from one of the websites we manage through PilotDeck.
They joined us early in beta, and within 3โ4 months, we started seeing strong organic progress without paid marketing.
Good early signal.
I still like Claude Code.
I am happy to pay $200 if the value is there.
I just want to understand what exactly $200 buys me.
Does anyone using the Max plan actually know how to predict their weekly usage? Can anyone at @AnthropicAI respond?
Paying $200/month for Claude and I still have no idea what I am actually buying. Am I the only one who does not understand this gimmicky Claude behaviour?
Some weeks we use Claude Code heavily and stay under 50%.
Other weeks we hit 100% ridiculously fast.
This week: 100% used.
And because of this we have started routing some coding work through GLM-5 behind the scenes.
The results have been surprisingly good.
And it is dirt cheap compared to what we are spending on Claude.