How long does it take a new page to start ranking?

In this case, about three weeks. Both of the pages that ended up driving most of the traffic were indexed within days but sat at position 8 to 10 with a handful of impressions for roughly 21 days before Google started showing them for the queries they now rank for. The content didn't change in that window; Google's willingness to display it did.

Do competitor alternatives and versus pages actually work?

Yes. Seventeen alternatives and versus pages produced about 1,660 visits in 28 days, all ranking on page one for a competitor's brand name, and they convert to App Store clicks two to three times better than informational posts.

Does ChatGPT send real traffic?

Some, and it converts. ChatGPT sent about 100 visitors in the last 30 days, the 8th largest acquisition channel, and 9% of them clicked through to the App Store, which is close to the rate for Google organic. The pages ChatGPT cites are the same pages that rank on Google.

Why didn't 175 programmatic pages produce more installs?

They produce impressions and visits, but the people landing on them are researching a specific settlement, not looking for an app. Most of those queries are answered in Google's AI Overview, so the click never happens, and when it does the visitor rarely converts. The pages that convert are the ones that sit in the buying decision: best-app listicles and competitor comparisons.

Mobile App SEO Case Study: 5x Revenue in 83 Days

On June 1st I started working with a mobile app that had done 500k downloads. They had invested nothing in SEO/AEO. 83 days later, web-attributed App Store revenue went from $433 a month to roughly $3,000, and two of the 294 pages I shipped did most of the work. Here are the numbers.

The numbers

June: $433
July: $1,833
August: on track for ~$3,000

How we attributed revenue

The revenue figures come from App Store Connect analytics. We set up a custom product page and pointed every download link on the website at it, so this is iOS revenue attributed to the website specifically. It does not include Google Play revenue, and it does not include installs that came from the App Store's own search.

Where we began

In early June I was handed a website with five live URLs: a home page, a blog index, privacy, support, and terms of service. That was all.

~88 website visits per day
~1,100 appearances in Google search results per day

What I did over the 83 days

Once I got GitHub repo access, in order:

Ran the site through Screaming Frog and cleaned up the technicals
Wrapped the site in GA4 and Google Search Console
Set up the SEO/AEO technicals: sitemap.xml, allowing AI bots, server-side rendering, schema
Created template pages for blog posts and for programmatic pages
Built a keyword and prompt set of 300 targets
Against those targets, shipped programmatic pages, alternatives pages, versus pages, and listicles

Totals: created 294 new pages (175 of them programmatic), refreshed 47, and deleted 9.

Since then, average website visits per day have gone up about 8x, from ~88 to ~650, and search impressions have gone up about 14x, from ~1,100 to ~15,000 per day.

What worked

One listicle is driving more organic traffic than the home page. We rank #1 for a listicle that reviews every competitor in our niche. It took about three weeks from publish to take off, and it now sends more people to the site than the home page does.

Alternatives and versus pages are cooking. There are 17 of them. They account for about 1,660 visits in the last 28 days, and every one of them ranks on page one for someone else's brand name.

ChatGPT visitors click through to the App Store at 9%. The same pages driving most of our Google traffic are also the ones LLMs cite most. ChatGPT sent about 100 people in the last thirty days. Small number, but it converts like search, and we expect it to be much higher next quarter.

Three pages outperform everything else. The 175 programmatic pages, combined, sit below the top three.

What failed

Volume was a waste. I shipped 294 pages. Two of them, one listicle and one post, sent about 200 people to the App Store in August. The other 292 combined sent fewer.

Broad prompts and informational pages were a waste. Informational pages get nice impressions and nice visit counts, but they convert at a fraction of the listicle and the alternatives pages. Most people read the AI Overview for those queries and never hit our site at all.

What we're prioritizing now

LLMs, via Bing and Brave. Most AI assistants pull from the Bing and Brave indexes, so ranking there is upstream of ranking inside ChatGPT, Copilot, and Perplexity. We've started to crack Google, so months 3, 4, and 5 are about those two indexes.

Reddit. I spent $500 on a Reddit consulting call. No data yet. I'll write it up once there is.

YouTube. AI Overviews love YouTube videos. Optimizing titles, descriptions, and captions for every client is going to matter.

If I could go back to June 1st

I'd skip straight to the programmatic engine, the alternatives pages, and the versus pages. Find the ~50 keywords and prompts that actually make money, get into the sources cited in those answers, and go very, very hard after them. That's the whole game. It's not pushing 300 pages, and I say that as someone who shipped 294.

Final thoughts

I treated SEO/AEO like a volume game. It's not. It's much more delicate. Two pages sent more people to the App Store than the other 292 combined.

The winners were buying pages: a listicle and competitor versus/alternatives URLs. The informational pages got impressions but did not get installs, because most of those queries never left the AI Overview.

The better bet is narrower and wider at the same time. Narrow the list to the 50 keywords and prompts that sit in the buying process. Then cover those 50 on-site and off-site: your page, the Reddit thread, the YouTube result, the random PR site ChatGPT already uses. Fifty buying queries, owned in more than one place, beat 300 mixed queries that only exist on your domain.

$433 to $1,833 to ~$3,000 in 90 days. The lesson wasn't doing more. It was deleting most of the plan and going hard on the part that pays.

Hopefully this is useful. Feel free to DM me any questions.

How We 5x'd Organic Search Revenue for a Mobile App in 83 Days