Chapter 11 · Every screen, part by part
Analytics
The same numbers as the dashboard, with room to ask questions of them. Every figure here traces back to real records, and clicking one opens them.
1 — How to read these. Opens the rules the whole page follows: what counts as researched, when a rate is suppressed for a small sample, and which sections the filters apply to.
2 — The filters. University, priority, degree type, and a date range. They apply to everything above What you are working with, and to nothing below it — that section says so in its own subtitle, because a lifetime total that silently obeyed a two-week filter would be the most misleading number on the page.
3 — Activity over time. Daily record counts across research, outreach and applications. The legend's number is the total across the window, not the last day.
4 — The window. 14, 30 or 90 days. This is separate from the date filter above and narrower: it changes the chart, not the page.
The conversion funnel#
Drop-off from researched through to an offer, each stage a link into the records behind it.
A funnel that goes up is telling you something. If a stage reads more than above, the chart will say which one and why underneath it. The usual cause is the publish gate: a match can be scored against a professor whose findings are still waiting in Review, and a professor only counts as researched once those findings are published. So Matched can exceed Researched, the numbers are both right, and publishing what is waiting brings them back in line.
A stage under 25% of the one above it is highlighted, because that is where the pipeline is actually leaking.
What you are working with#
Lifetime totals across the whole workspace, unfiltered and deliberately so. This is the section to read when you want to know how big the thing you have built is, rather than what happened this fortnight.
Reading rates honestly#
Two rules, applied everywhere:
- A rate always shows its denominator. "40%" of five records and "40%" of five hundred are not the same claim, and the page never lets you mistake one for the other.
- Too small a sample prints no rate at all. You get the counts and a note, rather than a confident-looking percentage built on three records.
Chapter 7 covers what to do with all of it, including a weekly rhythm that works.