"How do we know if the new website is actually working" is a question we get asked at the three-month mark on nearly every project. The honest answer requires deciding, before launch, which numbers actually matter for your business — because website analytics will hand you dozens of metrics that feel important and mostly aren't.
The KPIs that correlate with actual business outcomes
Conversion rate by traffic source, not just overall. A blended conversion rate hides the real story. We've had clients with a healthy-looking 3% overall conversion rate where organic search converts at 5.5% and paid social converts at 0.4% — the second number means that ad spend is being wasted on a landing experience mismatched to the audience, something the blended number never shows.
Cost per qualified lead, not cost per lead. Any form submission counts as a lead in most analytics setups. Cross-referencing form submissions against CRM data on which leads actually became sales conversations tells you the number that matters. We set this up with a closed-loop reporting connection between the website form and the client's CRM (HubSpot, Pipedrive, or a custom Zapier/Make automation) wherever budget allows.
Core Web Vitals as a leading indicator. LCP, INP, and CLS aren't just SEO factors — they correlate directly with conversion rate. Google's own research and data from our client sites both show conversion dropping measurably for every additional second of load time past 2-3 seconds. We track this monthly via PageSpeed Insights field data, not just at launch.
Task completion rate for the primary action. If the site's job is to generate quote requests, track the percentage of visitors who reach the quote form versus the percentage who actually submit it. A gap here usually means the form is too long, asks for information too early, or has unclear next steps — all fixable without a redesign.
A KPI dashboard structure that works
| Category | Metric | Review cadence |
|---|---|---|
| Traffic quality | Sessions by source, bounce rate by landing page | Weekly |
| Conversion | Conversion rate by source, form completion rate | Weekly |
| Lead quality | Cost per qualified lead, lead-to-sale rate | Monthly |
| Technical health | Core Web Vitals, uptime, 404 errors | Monthly |
| Revenue impact | Revenue attributed to site (e-commerce) or pipeline value (lead gen) | Quarterly |
Setting up measurement before launch, not after
GA4 event tracking needs to be configured for the specific conversion actions on the new site (form submits, phone clicks, quote requests) before launch, not bolted on after someone asks for a report. We build a measurement plan during the discovery phase alongside the site build, so the day the new WordPress site goes live, tracking is already validated against a staging environment test.
A results example
A custom WordPress build for a B2B manufacturing client included a redesigned quote request flow (from a 12-field form to a 4-field form with progressive disclosure for details) and closed-loop CRM tracking. Baseline cost per qualified lead was $340 CAD; six months post-launch, with the same ad spend, it was $198 CAD — not because more people filled out the form, but because the ones who did were better matched to what the sales team could actually close.
The mistake to avoid
Don't measure vanity numbers your board or manager will ask about just because they're easy to report (total pageviews, social shares) instead of the ones tied to revenue. A spike in pageviews from a viral post that never converts is not the same as a slow, steady increase in qualified leads — pick the dashboard that reflects which one your business actually needs.
Want help figuring out which numbers actually matter for your site? Tell us your primary conversion goal and we'll map the right KPI set to it.
Related reading
Surrey, BC
A full-service digital agency working in WordPress, Drupal, Shopify, Webflow, React, and React Native. We partner with universities, governments, and growing brands to ship sites and products that hold up after launch.
