Mia Zheng
Hello guys — we're researching GEO right now. It's basically the upgraded version of SEO.
Our ICP has gradually shifted from searching on Google to asking ChatGPT and Perplexity. We're no longer putting most of our budget into Google Ads.
Ranking #1 on Google is no longer the goal. The goal is to get your product cited more often inside ChatGPT and other model answers. You need more AI to quote you.
Our ICP moved from Google to ChatGPT
This is the part that changes everything downstream. If your buyers aren't opening a search results page anymore, then every dollar you spend optimizing for that page is spent on a surface they don't look at.
So our goal now is: how do we get our product to appear more often inside ChatGPT answers, or inside other language models' answers?
The goal: from ranking to being cited
Ranking first on Google is no longer what matters. What matters is that more AI cites you.
That's a different job. Ranking is about matching a query. Being cited is about being in the sources the model can reach.
Why Reddit is the channel to build
On the question of which channels to target, I did some research today and want to share it with you: focus on building your presence on Reddit.
Why do I say that? Here we need to look at it from a more technical, or more commercial, angle. Try it yourself — every time you ask AI something, look at which platforms and which pages it searched.
Reddit signed official data licensing agreements with both Google and OpenAI (the 2024 deals). So Reddit content gets crawled into AI answers on a priority, stable basis. This isn't a soft metric like "community trust" — it's an indexing privilege guaranteed by contract.
Why X isn't the same bet
Since Musk took over X (Twitter), external access and crawlers have been tightened. Google's crawlers have a hard time reliably crawling content on X, so you almost never see X as a source inside Google AI Overview.
Perplexity can occasionally pull some X content, but it's far less stable than Reddit.
Licensed to Google and OpenAI. Crawled on a priority, stable basis. Shows up consistently as a cited source.
Tightened external access post-acquisition. Rarely appears in Google AI Overview. Perplexity pulls it occasionally, unreliably.
More like this?
I write about GTM, GEO, AI agents, and the messy middle of scaling a B2B product. New post every 2–3 days.
Read more posts →Hello guys,我们正在研究 GEO 这件事情,其实就是 SEO 的升级版。
因为我们的 ICP 已经逐渐从在 Google 搜索,转向问 ChatGPT、Perplexity。我们已经不再把大笔预算都放在 Google Ads 上了。
在 Google 排名第一已经不再重要了。现在的目标是,如何让自己的产品更多地出现在 ChatGPT 或其他语言模型的答案里 —— 也就是,你需要让更多的 AI 引用你。
我们的 ICP 从 Google 转向了 ChatGPT
这一点会改变后面所有的事情。如果你的买家已经不再打开搜索结果页了,那你花在优化那一页上的每一分钱,都花在了一个他们不看的地方。
所以我们现在的目标是:如何让自己的产品更多地出现在 ChatGPT、或者其他语言模型的答案里。
目标变了:从排名到被引用
排名在 Google 第一名已经不再重要了,而是你需要让更多的 AI 引用你。
这是两件不同的事。排名是关于匹配一个 query,被引用是关于你在不在模型能够到的信息源里。
为什么重点建设 Reddit
在瞄准的 AI 以及发言的渠道上,我今天调研了些信息想要分享给你们:重点建设自己在 Reddit 的发言。
为什么这么说?这里我们需要更加技术、或者从商业背景的角度去看待这件事情。你可以去试下,你每次问 AI 后,她都是搜的什么平台、什么网页。
Reddit 和 Google、OpenAI 都签了官方数据授权协议(2024 年那笔交易),所以 Reddit 内容会被优先、稳定地抓取进 AI 答案里 —— 这不是"社区信任"这种软指标,是有合同保证的索引特权。
为什么 X 不是同一个选择
X(Twitter)自马斯克接手后,对部分外部访问和爬虫收紧过。Google 的爬虫很难稳定抓取 X 上的内容,所以 Google AI Overview 里几乎看不到 X 作为信息来源。
Perplexity 偶尔能拉到一些 X 内容,但远不如 Reddit 稳定。
与 Google、OpenAI 签有官方数据授权。被优先、稳定抓取,能持续作为引用来源出现。
易主后收紧了外部访问。几乎不出现在 Google AI Overview 里。Perplexity 偶尔能拉到,但不稳定。