You’ve already ditched the spray-and-pray spreadsheet.You have a cleaned list of prospects, enriched with firmographic data, technographic signals, and maybe even a whisper of intent from crawling their latest job postings.
Building Simple Dashboards with Free Software
Datasette as a Lightweight SEO Dashboard Engine: Querying Crawl Logs in Real Time
For years, the standard SEO dashboard stack looked something like a Google Sheet feeding a Looker Studio report—functional, collaborative, and free, but limited by the spreadsheet’s row cap and sluggishness when you try to pivot a few hundred thousand URLs. Meanwhile, the enterprise crowd flexes with full-blown Elasticsearch clusters or Snowflake instances. There is a middle path, one that costs exactly zero dollars and runs on a laptop or a cheap VPS, yet delivers SQL-level query power, interactive filtering, and near-instant response times even with millions of rows. The tool is Datasette, an open-source Python application by Simon Willison that turns any SQLite database into a browsable, API-able, and fully interactive web dashboard.
Why should a savvy tech nerd care about SQLite for SEO data? Because almost everything we export from third-party tools—Screaming Frog crawl logs, Google Search Console CSV dumps, Ahrefs’ backlink lists, or server log files—is a flat file that can be ingested into a relational table with minimal friction. Datasette’s killer feature is that it does not require you to write a single line of front-end code. You point it at a SQLite file, and it instantly gives you a faceted browse interface, a full SQL query editor, and JSON endpoints for every query. For a startup marketer who is also the de facto data engineer, this means you can stop wrangling pivot tables in Sheets and start asking questions like “show me all pages with a crawl depth greater than 3 where the status code is 200 and the title tag is missing” in under a second.
The setup is brutally simple. Install Datasette via pip, then run `datasette mydata.db`. That’s it. Your local browser opens to a page where you can filter by column values, use faceted navigation, and export filtered results to CSV or JSON. To build a live dashboard, you need to automate the data pipeline: write a simple Python or Bash script that downloads your crawl log, transforms it into a normalized schema, and writes the rows into a SQLite database. Tools like `csvs-to-sqlite` or `sqlite-utils` (both by the same author) handle bulk ingestion cleanly. You can schedule this script via cron on a $5/month DigitalOcean droplet and deploy Datasette as a public-facing web app using something like `datasette publish fly`. Now your entire team has a shared dashboard that updates every hour with fresh crawl data.
The real power reveals itself when you start writing parameterized SQL queries and saving them as canned report views. Datasette lets you define custom SQL templates in a metadata file, which becomes a dropdown menu of reports on the homepage. For instance, you can create a query that joins your crawl data with a table of Google Search Console impressions, compute a “crawl efficiency” metric, and render it as a bar chart using the built-in Vega visualization plugin. The plugin system is modular: add `datasette-cluster-map` to map geographic origins of traffic, or `datasette-vega` for any chart type you can describe in Vega-Lite JSON. You are not locked into a rigid dashboard tool; you are composing a dashboard out of SQL queries and declarative visualizations, which is exactly the kind of control a seasoned marketer wants.
One obvious use case is monitoring page speed regressions over time. Export Lighthouse scores from a scheduled headless Chrome run (using free tools like Lighthouse CI or a GitHub Action), shove them into a Datasette database, and create a SQL view that calculates the moving average of CLS for your top 100 landing pages. Then add a `datasette-graphql` plugin to expose that view as a GraphQL endpoint, and your React-based custom dashboard (if you insist on over-engineering) can consume it. But you probably will not need to, because Datasette’s own UI is responsive and shareable by URL. Send your product manager a direct link to the filtered view of “all pages with FCP > 3s and traffic > 500 sessions last month” and they can interact with it without bugging you.
The trade-offs are real and worth acknowledging. Datasette is single-node and single-threaded by default, so it will not replace a proper data warehouse for multi-terabyte datasets. But for the typical startup SEO data set—a few gigabytes of crawl files and search console exports—SQLite performs admirably. You also lose the real-time streaming feel of something like Grafana, but a cron job every 15 minutes is often enough for strategic decisions. The bigger win is escaping the spreadsheet’s curse of accidentally broken formulas and rows that mysteriously vanish when you sort a column. Datasette treats data as immutable tables; you delete nothing, you filter everything.
If you have ever felt the urge to build a “quick and dirty” dashboard that ends up taking three days because you need to configure row-level security, custom CSS, and a login system, Datasette offers a saner path. It ships with OpenID Connect authentication via plugins, so you can lock the dashboard behind Google accounts in minutes. The startup marketer who learns to love SQLite and Datasette will find themselves with a superpower: the ability to turn any pile of CSV exports into a queryable, shareable, and visually rich dashboard without a single dollar of SaaS spend. That is the kind of leverage that scales from bootstrapped side-projects to Series A teams.
