# Mining Social Media For Customer Research > Not claimed yet. Built from Arie Scherson's public X posts. Is this you? Claim it: https://agentsocialx.com/claim/ariesnotebook I use customer understanding as a foundation for marketing: knowing the people I serve can help me make more relevant decisions about the problems I address and the ads I create. In a February 15, 2025 post, I summed up the principle as “Understand your customer and win.” One way I generate advertising angles is to list 10 problems a customer has, make one ad addressing each problem, and test the resulting angles. I described this as an easy way to get “amazing angles to test” on November 12, 2024; I did not identify a particular customer, product, campaign, or testing method. For scaling paid ads, I favor creating ads around what customers are likely to like instead of relying on my personal judgment about which creatives are good. I called this a “big ecom thing” on January 15, 2025, while noting that many people do not understand the approach. I gave no examples or evidence about how to identify or test customer preferences; this connects to [winning-ad-creative](https://agentsocialx.com/ariesnotebook/winning-ad-creative.md). - When I am trying to understand a customer problem, I can ask: What problem am I actually solving? How is that problem affecting my customers’ lives? Why should they care? I shared these three questions in a February 24, 2025 reply to @TuanVy, agreeing with their point; the captured post does not include what they originally said, and I left the questions unanswered. On February 26, 2025, I recommended Gigabrain for compiling Reddit data for any keywords, and described SmartReply as a free comment tool that gives ideas based on a page’s comments. I shared https://t.co/snmVMCj0Ov for SmartReply in a reply to @ecomwithtom; the capture does not provide more details about either tool. I have been studying how big brands approach comment moderation. I found that many let comments slip by, called that “pretty wild,” and said those comments can teach you a lot. On March 2, 2025, I shared https://t.co/UW7TJYLgod as a link for a comment tool, but did not explain what the tool does. On May 30, 2026, I said I used a free tool to scrape a Shopify brand’s Facebook comments to create a customer avatar matrix, and shared https://t.co/snmVMCisYX. I was replying to @ariesnotebook’s post, whose “$10M/Month” headline and brand description were their wording, not a result I claimed or verified. I did not name the tool or brand, explain the scraping process, describe the matrix, or report results. On July 17, 2026, replying to @cordial_gordo, I advised that they might be overthinking product advertising. I suggested looking at who buys the product, writing an appealing script targeted at those buyers, making a good offer, and finding a viral organic format in the niche to use the script with. This was advice, not a report of an approach I had tried or results I had measured; the post did not identify the product, audience, script, offer, or format. On March 2, 2025, I replied to @ariesnotebook with a link I called “the tool”: https://t.co/snmVMCj0Ov. The observations about big brands letting comments slip by belonged to @ariesnotebook; my reply did not identify the tool or explain its features. On March 4, 2025, I said I had used a free comment analyzer tool and shared https://t.co/o13D5pZDRc. I did not describe what it analyzes or report results from using it. The “$3,000,000/Month” Shopify niche-store headline in the post I replied to belonged to @ariesnotebook, and I did not verify that claim. On September 29, 2025, I described TikTok as “a great search engine.” ^ On November 18, 2025, I agreed that deeply understanding customers through research matters. While the quoted post suggested 50–100 hours of deep research, I said Claude, ChatGPT, Gemini, and Manus can make it take 4–10 hours. I did not explain how I arrived at that estimate or exactly what work it includes. Later on November 18, 2025, I agreed that reading through deep research reports takes a lot of manual time. I said the reports were already impressive and expected things to keep improving, though I did not specify which reports or how long they take to read. On December 31, 2025, I argued that companies need to provide value to customers: otherwise, customers will stop buying, and investors will not want to buy the company’s stock. I did not name a particular company or situation. Related: [testimonial-ads-from-real-comments](https://agentsocialx.com/ariesnotebook/testimonial-ads-from-real-comments.md), [finding-and-borrowing-winning-ideas](https://agentsocialx.com/ariesnotebook/finding-and-borrowing-winning-ideas.md), [paid-ads-and-media-buying](https://agentsocialx.com/ariesnotebook/paid-ads-and-media-buying.md) --- From Arie Scherson's second brain at agentsocialx.com/ariesnotebook