Waypoint

<!-- TODO: Add screenshot/video of Waypoint — show the pipeline funnel, SLA breach flags, and deal velocity charts -->

#Why I built this

The team reviewed the pipeline weekly but had no single view of how deals were moving. Stage conversion rates, time-in-stage, and SLA breaches were scattered across Airtable views and spreadsheets. The question at every meeting was the same: "which deals need attention right now?" I built Waypoint to answer that question in a 30-second glance.

#What it does

Waypoint tracks where every application stands in the investment pipeline. It shows stage-by-stage conversion rates, deal velocity metrics, and SLA flags — so the team knows how fast deals are moving and where bottlenecks form.

#Data model

#Pipeline stages

The pipeline follows a defined stage sequence, each with its own SLA:

Stage	SLA (days)	Description
Inbox	3	New application, awaiting initial review
First Call	7	Scheduled or awaiting first partner call
Final Interview	10	GP-level interview scheduled or in progress
IC Review	5	Investment committee deliberation
Term Sheet	14	Offer extended, awaiting founder response

SLAs are configurable per batch (each cohort has different velocity targets). A deal breaching its stage SLA gets flagged with a red indicator and surfaces at the top of the dashboard.

#Airtable sync

Waypoint pulls from the Applications table in Airtable via the REST API. The sync runs on page load with a 5-minute cache (Vercel KV). Each application record includes:

• Current stage and stage entry timestamp
• Partner assignments (inbox reviewer, first call partner, GP interviewer)
• Batch identifier and application metadata
• Stage transition history (computed from Airtable's Last Modified Time per stage field)

#Computed metrics

From raw stage data, Waypoint computes:

• Conversion rates — % of deals advancing between each stage pair, with batch-over-batch trend
• Velocity — median time-in-stage per stage, flagging outliers beyond 1.5× the median
• SLA compliance — % of deals currently within SLA per stage, with breach count
• Partner load — deals per partner at each stage, for workload balancing

#Dashboard design

#What got cut

The first version had 12 widgets. By the third pipeline review meeting, only three were being referenced:

• Cut: total deals ever (only goes up — nobody acts on it), source channel breakdown (useful quarterly, not weekly), geographic distribution (interesting but not actionable in a pipeline review), partner NPS scores (wrong meeting for this)
• Kept: stage conversion funnel, deal age distribution with SLA flags, partner load heatmap

The principle: if a widget doesn't change how someone acts after the meeting, it's noise.

#SLA as primary alert

Rather than color-coding every metric on a gradient (green → yellow → red), Waypoint uses binary SLA flags. A deal either needs attention or it doesn't. Defining "too long" per stage turned out to be more valuable than the dashboard widget itself — it forced the team to articulate what "healthy velocity" actually means.

#Technical decisions

• Vercel KV caching over real-time sync — Airtable's API has aggressive rate limits (5 req/sec). Caching with a 5-minute TTL gives near-real-time data without hitting limits during a meeting where 8 people load the dashboard simultaneously.
• Server components over client-side fetching — all data fetching happens in Next.js server components. The client receives pre-rendered HTML with the latest metrics. No loading spinners, no waterfall fetches.
• No charting library — the funnel and heatmap are CSS Grid + inline styles. Recharts or Chart.js would have been faster to build but harder to match the exact visual density the team wanted. Custom rendering gave full control over pixel-level layout.

#Stack

Next.js 15, React 19, Airtable API, Vercel KV (caching), Tailwind CSS, CSS Grid (custom charts), Vercel