Maxwell

Indie maker building AI + e-commerce tools. Failed a startup, learned more from it than success. Shipping with vibe coding, one prompt at a time.

[Maxwell @max_mtwx]

@DavisPaipa yeah. people obsess over ranking on ChatGPT and forget Brave, Perplexity, Copilot all pull from the same pages. optimize the page, not the model.

[Maxwell @max_mtwx]

@vladislavgeras the em-dash tell is real. been called out for it twice. now I strip them before anything ships. in-browser so text never leaves is the part everyone else gets wrong.

[Maxwell @max_mtwx]

@OderaOfficial the post-checkout 'wait that address is wrong' panic is real. usually turns into a support ticket and a refund. auto recalc kills it before the buyer even notices. more useful than half their splashy feature drops.

[Maxwell @max_mtwx]

@ecomchai check the box size before the rate. if the box is oversized at 1.3kg you're paying dimensional weight, and shrinking it cuts cost with zero negotiation. and if shipping is more than the product, the product is too cheap, bundle to lift AOV instead of fighting the freight.

[Maxwell @max_mtwx]

@vladislavgeras this one is sneaky because it passes the rich results test. the validator checks each schema in isolation, it never flags two of them disagreeing. that's exactly why it ships to production.

[Maxwell @max_mtwx]

The underrated line: lowest-effort Fable 5.1 beats max-effort Fable 5. Effort level just stopped being a hidden knob and became a real cost dial — you don't need max for everything anymore, and now there's a price curve to prove it.

[Claude @claudeai]

As well as being capable of much higher performance than Fable 5, it can also achieve similar or better results at a much lower cost when set to lower effort levels.

[Maxwell @max_mtwx]

@KaiCromwell DR is a lagging proxy, not a ranking factor, so I mostly agree. a 0.9 DR site ranking for transactional terms still has a few real links to that exact page, they just don't move the DR needle. people blame DR instead of checking on-page and intent match.

[Maxwell @max_mtwx]

@CarlWeische listicles are doing way more of this than the media buying crowd admits. TOF gets cold traffic in, the listicle is what closes the person who was never buying off an ad. skip it and TOF looks broken.

[Maxwell @max_mtwx]

@ecombender and it pads every report so you think it's crushing. exclude your brand day one or you're just paying for demand you already had.

[Maxwell @max_mtwx]

@samigrows less than I should. research is quick, mapping eats the hours. cannibalization is only real once pages are live and competing for the same term.

[Maxwell @max_mtwx]

@EcomShri fair, AC has a shoulder season so there's still volume to work with. the cope test is a truly dead season like holiday decor in March, where the search just isn't there no matter the creative

[Maxwell @max_mtwx]

@ecombender which one? a lot of these measure the product but never push the numbers into your size chart or review photos, and that's the bit that actually cuts returns

[Maxwell @max_mtwx]

@EcomShri agree for non-brand and shopping. for a branded search campaign the name IS the keyword, so leading with it is fine. the rule needs that split.

[Maxwell @max_mtwx]

@wahg_one a whole day on one product page is the part the streak hides. what finally got it to click?

[Maxwell @max_mtwx]

@CarlWeische the hidden no-thanks doesn't just cost the upsell, it costs the next order. people remember the pushy PPU moment and hesitate at checkout next time. trust compounds both ways.

[Maxwell @max_mtwx]

if you're building backlinks, read this before you chase DR. the indexing check is where founders lose the game. from the directory side I can confirm, a listing on an uncrawled page is a backlink worth zero.

[Sami | SEO AI @samigrows]

A DR 90 backlink is worthless if Google never indexes the page it sits on.

This is where most founders lose the game.

What kills indexing:

Directory page is noindexed
Directory page has zero internal links
Directory hasn't been crawled in months

What guarantees indexing:

Check the directory's own sitemap first
Submit through IndexNow after listing goes live
Pick directories with active blog sections (proof of crawl activity)

Crunchbase, G2, and Product Hunt pass all three checks. That's why they show up in every high-DR backlink list.

Speed means nothing if the link sits unindexed for months.

Reply "INDEX" for the indexing checklist.

[Maxwell @max_mtwx]

@samigrows flip side, I run a directory so I'm the one who has to get the listing pages indexed. thousands of near-duplicate pages and no blog section on each. do internal links from category hubs carry enough, or is thinning the catalog the only honest fix?

[Maxwell @max_mtwx]

@ecombender optimization score is just a feature adoption meter. I read ROAS and contribution margin, the platform reads its own revenue.

[Maxwell @max_mtwx]

@ecomchai still the same avatar. avatar is the who, angle is the hook. pain vs social proof are two angles for one person. test avatars first means list your customer segments, pick the one with the sharpest problem, then run angles inside it

[Maxwell @max_mtwx]

the line in the video that actually explains the mechanism is your headline. ripped ads convert because the mechanism is obvious, not because of editing. put that sentence on top of the product or a before/after frame, done. don't try to compress the whole video into one image, just say the one thing that matters