reddit · SideProject · 10-30 06:50 42 分 ← 返回列表

I built a tool that turns earnings reports into Sankey diagrams—but data automation is the bottleneck. Should I pivot or quit?

I’ve been building a side project called Sankey Earnings. It turns company earnings data into interactive Sankey diagrams showing how revenue flows through costs, operating expenses, and profit. People can use it to: Quickly understand a company’s quarterly financial performance Explore different companies and reporting periods Read a concise earnings summary alongside the visualization Export charts as PNG or SVG for research, presentations, or social media I’ve also built an

五维评分

付费意愿

需求真实度

个人可落地

竞争空间

新鲜度

AI 分析

痛点：财报数据格式高度不统一（PDF、新闻稿、IR 页面、表格标签各异），AI 抽取结果不准确，错误代价大，导致财报结构化与可视化产品无法规模化自动化。

目标用户：独立投资者、股票研究分析师、财经自媒体或内容创作者，以及需要白标财报图表的财经媒体与研究机构。

证据：The problem is data ingestion. Earnings reports aren't standardized enough for reliable automation... Even with AI-assisted extraction, financial accuracy matters too much to blindly publish the result. Consequently, the product still needs substantial handholding: configuring companies, checking extracted numbers, fixing mappings, and reviewing charts before publishing.

变现：订阅制 SaaS：按覆盖公司数量、导出次数或 API 调用量分层收费；面向财经媒体或研究机构提供白标图表和结构化财务数据 API；免费精选页面配合付费导出和高级公司覆盖。

MVP：两周内只覆盖 10 到 20 家美股热门公司，用半自动流程（AI 抽取加人工校验模板）生成精选 Sankey 财报页，配邮件订阅和 PNG 与 SVG 导出付费墙，先验证是否为导出或覆盖更多公司付费，不做全自动数据管道。

竞品：Koyfin、FinChat、Bamsec、Simply Wall St、Stratosphere.io、AlphaSense

摘要：作者做了一个把公司财报转成交互式桑基图的 side project，可导出 PNG 或 SVG 并配财报摘要，但核心瓶颈是数据接入：各公司财报表格、PDF、IR 页面结构差异极大，即便用 AI 抽取也需要大量人工核对映射与图表，扩展覆盖极其费力。他正在纠结是收缩为少量热门公司的精选网站、把可视化导出工具卖给金融从业者，还是直接放弃。帖中没有任何付费用户或收入数据，需求仍停留在作者自述层面。