Kai Jiang

Solo dev building & shipping products in public.
First launches rolling out now →

[Kai Jiang @jqz_dev]

@ankittharol I'd put the launch directories money last. Most of the free ones do the same job, and the useful part is checking a week later which ones sent real clicks. Listings in a niche category tend to keep sending traffic long after launch day.

[Kai Jiang @jqz_dev]

@MaberFate "Useful data competitors don't have" is the part doing most of the work. Ten pages with something original beat a hundred rewrites of the current top results, and they tend to pick up links on their own.

[Kai Jiang @jqz_dev]

@tirthaexe Mostly agree. The exception I'd make is when the product itself is what people search for, like a free tool or calculator. Then one good page can bring users before you have a clear ICP, because the intent is already there.

[Kai Jiang @jqz_dev]

@seosmarty The gap between what Google says and what gets sold is wild. The sites that seem to hold up use AI for drafts and data, then have someone who knows the topic decide what goes on the page. Publishing straight from the model at volume is what keeps getting hit.

[Kai Jiang @jqz_dev]

@foliofed Not starting one right now, but curious what trips people up most on indexing. What I've noticed is that template pages with data that actually differs page to page get indexed, and the ones stuck in "crawled, not indexed" are usually near duplicates of each other.

[Kai Jiang @jqz_dev]

That's the trap. Founder communities are great for feedback, but most people there are selling, not buying. The question I'd ask for each of those five is where their actual users go when they have the problem, whether that's Google, a subreddit or an app store. That's usually the channel worth the month of effort.

[Kai Jiang @jqz_dev]

@Graphed One of the cheapest wins there is. One caveat: if the query has a different intent than the page, adding it can water the page down. A quick look at what currently ranks for it tells you whether it belongs as a section or deserves its own page.

[Kai Jiang @jqz_dev]

@AleksejAros Seeing this a lot lately. Splitting by query usually explains it: a few head terms lost clicks to an AI Overview or a new SERP feature while the long tail held steady, and the average hides both. Checking CTR per query first saves a lot of pointless title rewrites.

[Kai Jiang @jqz_dev]

@sakansokan Worth checking each directory's own traffic before spending time on it, plenty of high DR ones get almost no real visitors. The small niche ones where people actually browse a category have sent me more clicks than the big generic lists.

[Kai Jiang @jqz_dev]

For a small site, the three that did the most for me: pages built around the exact question people type instead of the broad keyword, refreshing pages already sitting around positions 8 to 20 in Search Console, and internal links from pages Google already crawls a lot. Backlinks mattered, just less than I expected early on.

[Kai Jiang @jqz_dev]

@iniyanai Nice, 2.4% from directory traffic is better than most people get from a launch day. Worth tagging which directory each sale came from, because it's common for one or two of them to do almost all of the work while the rest send a trickle.

[Kai Jiang @jqz_dev]

@NamanBansal0611 The ChatGPT app showing up as direct is the one people miss most. A quick sanity check is to compare direct traffic on deep pages against the homepage, since nobody types a long blog URL by hand. A jump there usually means AI referrals with the referrer stripped.

[Kai Jiang @jqz_dev]

@David_cbolt Point 3 is the underrated one. A page sitting at position 6 already has Google's trust, so a clearer first two lines can move it more than a brand new post would. How long did it usually take to see the CTR change after you rewrote one?

[Kai Jiang @jqz_dev]

@fabiobergmann YouTube at the top lines up with what I've seen, especially for Perplexity. Did the split change a lot by engine? Averaging across all of them can hide where the effort actually pays off.

[Kai Jiang @jqz_dev]

@Nick_zv_ Number 6 saves so much wasted writing. I'd run the same SERP overlap check on pages that already exist too, since two posts sharing most of their top 10 are usually splitting clicks. Merging them often helps more than writing the next one.

[Kai Jiang @jqz_dev]

The thin content one catches a lot of programmatic sites that never thought of themselves as spammy. A thousand pages with the same template and one swapped variable reads as doorway pages to a rater. Have you seen many recoveries where pruning alone was enough, or do they usually need a rebuild?

[Kai Jiang @jqz_dev]

@goaimode Agree. The comparison step now happens inside the answer, so being mentioned on the pages the model reads, like reviews, roundups and Reddit threads, matters as much as your own site. Your site still has to answer the follow-up questions clearly once someone does click.

[Kai Jiang @jqz_dev]

@thisaliraza Painful but common. Before deleting anything I'd export every page with clicks in Search Console over the last 12 months, and anything with traffic or backlinks gets a 301 to the closest match. Seasonal pages especially, since they look dead in a summer audit.

[Kai Jiang @jqz_dev]

@hellofoundryai Interesting split. Makes sense that AI answers lean on reviews and on-page content, since that's text a model can actually quote, while profile signals mostly feed Maps directly. Did the survey say anything about mentions in local roundups or news?

[Kai Jiang @jqz_dev]

@jaydippatel83 Boring tools win because the big sites never bother making a dedicated page for them. Curious whether the ICO page sends people on to your other tools, or if most leave after one conversion.