A 60-Day Review Plan for a New Graph Learning Framework Website
Launching a technical documentation website for a library like TGraphX is
not a single event — it is the start of an iteration cycle. The first 60 days
provide data on what's working, what's missing, and what needs revision. This
plan covers what to measure, when to measure it, and how to act on the results.
What the 60-Day Window Captures
A new site takes time to be indexed, ranked, and visited. The 60-day window
is long enough to see initial indexing signals but short enough to make
meaningful course corrections before the content calendar gets too far ahead.
After 60 days, you can answer:
- Is the site being indexed by search engines?
- Which queries are bringing visitors?
- Are visitors finding what they came for (or bouncing immediately)?
- Is there organic interest in the library's specific features?
- Which articles are getting traction and which aren't?
You cannot answer after 60 days:
- Whether a long-term SEO strategy is working (that takes 6-12 months)
- Whether content is driving library adoption (depends on trial-to-adoption lag)
Measurement Framework
1. Index Coverage
What: Are search engines finding and indexing your pages?
How: Google Search Console → Coverage → Indexed pages count. Compare
to submitted sitemap URL count.
Target by day 60: All submitted pages should be indexed or in queue.
Unindexed pages after 45+ days indicate crawlability issues (robots.txt,
noindex tags, slow load, thin content).
Action if missing: Check for noindex tags, canonical URL conflicts,
slow server response times.
2. Search Queries and Impressions
What: Which search queries are showing your pages, and how often?
How: Search Console → Performance → Queries. Sort by impressions.
What to look for at 60 days:
- Branded queries ("TGraphX", "tgraphx python", "tgraphx tutorial") — these
should appear early as people look up the library by name
- Feature queries ("tensor-valued graph neural network", "graph shape
validation") — these indicate organic discovery
- Comparison queries ("TGraphX vs PyG", "TGraphX vs DGL") — these indicate
users evaluating the library
Action if no queries: Content may be too thin, too similar to existing
content, or the library may not yet have enough name recognition to drive
searches.
3. Click-Through Rate (CTR)
What: When your pages appear in search results, what fraction of users
click?
How: Search Console → Performance → by page or query. CTR column.
Benchmarks: Technical documentation typically sees 3-8% CTR from
informational queries. Branded queries (people searching for "TGraphX") see
higher CTR (15-30%) because searchers know what they want.
Action if CTR is low: Review title and meta description for pages with
high impressions but low CTR. The title/description is the ad for the page.
4. Article Quality Signals
What: Are users reading the articles or bouncing?
How: Google Analytics or Plausible → Engagement rate, scroll depth,
time on page.
Targets:
- Average time on page for 1000-1800 word articles: 3+ minutes
- Scroll depth: 50%+ median
- Bounce rate (single-page sessions): below 70% for tutorial articles
Action if bounce rate is high: Check whether the page answers what the
title promises. A title mismatch (promises a tutorial, delivers an overview)
drives immediate bounces.
60-Day Review Table
┌────────────────┬─────────────────────────────┬──────────────────────────┐
│ Day │ Measurement │ Action threshold │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 7 │ Crawl confirmed (Search │ If not crawled: check │
│ │ Console URL inspection) │ robots.txt, sitemap.xml │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 14 │ Index count in Search │ If 0 pages indexed: │
│ │ Console │ investigate indexability │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 21 │ First branded query │ If no branded queries: │
│ │ impressions │ library name may not │
│ │ │ have enough recognition │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 30 │ Article-level impressions, │ Flag articles with 0 │
│ │ page coverage │ impressions at 30 days │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 45 │ CTR by article, top queries │ Revise titles with │
│ │ │ CTR < 1% and 100+ │
│ │ │ impressions │
├────────────────┼─────────────────────────────┼──────────────────────────┤
│ Day 60 │ Full review: index, │ Produce 60-day report; │
│ │ queries, CTR, engagement, │ prioritize next 30 days │
│ │ article quality signals │ of content │
└────────────────┴─────────────────────────────┴──────────────────────────┘
What to Measure for a Technical Library Site Specifically
General website metrics apply, but technical library sites have specific
signals:
GitHub traffic correlation: Check if website traffic spikes correspond
to GitHub star events or releases. A release post on the website should
show a measurable traffic increase.
Documentation completeness queries: Search Console queries containing
"how to", "example", "tutorial" indicate users looking for documentation
that may be missing. These are content gap signals.
Comparison query performance: Queries like "TGraphX vs PyG" have high
commercial intent — the user is evaluating tools. Articles addressing these
comparisons directly ("TGraphX vs DGL: Tensor-Valued Graph Workflows") should
perform better than generic comparison articles.
Error and FAQ queries: Queries containing "error", "not working", "fix",
"shape" indicate users with specific problems. FAQ articles and error
documentation address these directly.
Content Quality Review at 60 Days
Beyond traffic metrics, review content quality by asking:
-
Accuracy: Do any articles make claims that the TGraphX source does
not support? (Use the source-grounded verification approach.) -
Completeness: Are there common use cases not covered? Check GitHub
issues and Stack Overflow for unanswered TGraphX questions. -
Freshness: Has TGraphX released any updates? Do articles reference the
correct version (1.4.2)? -
Code correctness: Run code examples in articles against the current
TGraphX version. Deprecated APIs or changed function signatures should
be updated.
The Iteration Decision
At day 60, make three decisions:
Continue: Articles driving traffic with good engagement — keep producing
similar content.
Improve: Articles indexed but with poor CTR or high bounce — revise titles,
improve introductions, add diagrams.
Drop or redirect: Articles with no impressions after 60 days and no clear
keyword opportunity — redirect traffic to a more focused piece.
The 60-day review is not about declaring success or failure. It is about
having enough data to make confident decisions about the next 60 days of
content and technical work.