Original Summary

I spend more than a year building products, especially one product (launched May 2025) where I'm spending every tiny detail to improve the platform, remove friction and optimizing the funnel. The MRR and revenue keeps growing last 6 months, and it's not because I was building. It was because I did research first. Most of it told me I had been optimizing for the wrong search engine. 1. I checked which engine my buyers actually use. Everyone defaults to Google. My search engine signups over a 60 day window: Bing: 744 DuckDuckGo: 426 Google: 58 So I started reading Bing Webmaster Tools as seriously as Search Console, and pinged IndexNow on every content change (the Bing family consumes it, Google does not). Last 30 days: Bing 5,871 clicks from 36,413 impressions. Google 1,973 from 22,651. Bing still sends about 3x. 2. Google was not ignoring my content. It could not find it. Sitemap not read since April The money page unknown to Google The whole SEO estate orphaned from the footer, nothing internal pointed at it Fixed the footer links, made sitemap lastmod honest instead of decorative. 3. The on page work was boring. 13 titles at 66 to 75 characters, truncating in the SERP, trimmed under 60 A double brand name bug on two pages A hardcoded schema dateModified frozen at one date sitewide Blog articles taught to link the landing pages 4. SEO pages beat blog articles, and it was not close. One seo landing page: 633 signups in 60 days, 24% of everything I got. The entire blog in the same window: about 48. So I cloned the page pattern instead of publishing more articles. 5. E-E-A-T is not a score you add. It is a list of things you refuse to fake. The rules I write under: Real organizational author by default. A named person only when they actually own the argument Never invent a person, bio, credential, test, customer, testimonial or statistic. A configured author bio in your CMS is not proof that a person exists Every claim matches shipped behaviour, not the roadmap Every competitor fact gets a source URL and a last checked date, visible on the page My own comparisons disclose that I own the product, and give a real case where the competitor fits better A real corrections and contact route exists before anything publishes None of that is a ranking trick. It is what stops you publishing something you have to walk back. 6. AI search: three metrics, none of them traffic. Bing AI citations, last 30 days: about 11,800, roughly flat on the prior period. About 96% of them point at one single URL Google generative AI impressions, last 90 days: about 2,440 Visitors actually arriving from AI channels in 30 days: 113 Against 5,416 organic search visitors in that same window. So I track AI visibility, but I do not build for it. No llms.txt as a tactic, no special AI schema (there is none), no near duplicate pages for conversational query variants, no promising anyone a citation. A citation is visibility, not authority and not a sale. Where it sits, 90 days

  • 情报分类：商业与市场研究
  • 分类依据：内容涉及商业、投资或市场动态
  • 信息来源：Reddit · SaaS
  • 发布时间：2026/9/18 17:44:22