A 20-minute weekly SaaS performance review

A weekly review should reduce uncertainty, not create a tour of every available metric. Twenty focused minutes is enough when the definitions are stable and the evidence is already connected.

This routine uses Revinho's current Overview, Growth, Funnels, Pages, Revenue, and Opportunities surfaces. It is a manual operating workflow; it does not depend on an automated weekly email.

Before the timer starts

Choose one reporting period and keep it consistent across the review. For a weekly rhythm, use the most recent seven complete days compared with the preceding seven complete days in the workspace timezone.

Avoid including a partial current day. Check provider freshness and connection health before interpreting a sudden drop. Write down material context such as a launch, outage, pricing change, campaign, or content update as an annotation—but remember that an annotation is context, not proof of cause.

The meeting needs one person who will record the decision. It does not need a large audience.

Minutes 0-4: read the business outcome

Open Overview and answer only:

• Did net revenue rise or fall?
• Did current MRR expand or contract?
• How did new customers and new MRR change?
• Did sessions, users, and key events move in the same direction?

Read both the absolute and percentage changes. A large percentage on a tiny base should not dominate the review.

Then scan the ranked change signals. Each useful signal should name the period, current and previous values, and the deterministic rule that produced it.

Do not explain the numbers yet. Pick the one movement with the greatest combination of magnitude, commercial relevance, and evidence quality.

Minutes 4-8: locate the acquisition movement

Open Growth and break sessions down by channel, source, medium, and campaign.

Ask:

• Which segment contributed most of the absolute change?
• Did its key events move with its sessions?
• Did the mix shift between high-intent and informational traffic?
• Is the result large enough to interpret, or is it a small-base percentage spike?

If Search Console is connected, inspect organic queries and page demand. Search Console impressions and clicks describe Google Search visibility; they do not represent total market search volume.

If Stripe revenue moved during the same period, describe it as associated context unless a deterministic join exists.

Minutes 8-11: inspect the funnel

Use the saved or preset GA4 funnel closest to the product's real journey.

Read:

• starters and completions;
• overall completion;
• the largest exit;
• step-by-step retention;
• sampling or unavailable states.

The funnel is based on events already configured in the selected GA4 property. If the “purchase” event is missing or incorrectly fired, the funnel describes that instrumentation—not necessarily every real Stripe purchase.

Choose the largest meaningful exit, not the visually largest bar. A drop from 1,000 page views to 20 purchases may be normal for a broad landing page; a sudden week-over-week change concentrated on one device is more actionable.

Minutes 11-14: find the page that deserves work

Open Pages and look for one of four evidence patterns:

• high traffic with weak key-event behavior;
• a hidden winner with strong behavior and limited traffic;
• a fast-growing page whose next step needs review;
• a declining or striking-distance page with meaningful search demand.

Inspect the page profile rather than the label alone. Compare GA4 sessions and key events with Search Console impressions, clicks, CTR, position, and top query where available.

A page with growing impressions and stable position may be benefiting from increased demand. A page with stable impressions and falling CTR needs a different response. When organic clicks rise while key events stay flat, trace the page and query mix before assuming the traffic is unqualified.

Minutes 14-17: reconcile the commercial detail

Open Revenue and separate:

• gross revenue;
• refunds;
• net revenue;
• current, new, and churned MRR;
• new and existing customer mix;
• subscription and one-off streams;
• product and plan mix;
• native and converted currency context.

This prevents an acquisition story from hiding a billing explanation. Revenue might be lower because a large renewal moved outside the window. MRR might be lower because an existing subscription churned while new-customer count improved. Refunds might erase stable gross payments.

Use the Revinho methodology whenever the team disagrees about a definition. Do not resolve a metric disagreement by changing the definition for the week.

Minutes 17-20: choose one action

Open Opportunities and review the evidence for the highest-priority candidate. Then choose one action that is small enough to complete before the next review.

Good actions are specific:

• Verify the mobile checkout after a conversion-rate decline concentrated on mobile.
• Improve the next step on one high-traffic, low-key-event landing page.
• Refresh one page whose clicks and position declined across equivalent periods.
• Contact the customers behind the largest churn movements.
• Correct a duplicated purchase event before interpreting GA4 revenue.

Record:

Observation: what changed, with current and previous values
Evidence: the source, period, segment, and provenance
Action: one change or investigation
Owner: one person
Review date: the next weekly review
Expected evidence: the metric or behavior that would move

Do not assign three unrelated projects from one signal. The purpose of the review is focus.

A worked weekly summary

A defensible summary might read:

Net revenue declined 12% while sessions increased 18% during the same seven-day period. The session growth was concentrated on two organic guides, while the visitor-to-key-event rate fell from 4.1% to 3.2%. Stripe refunds were unchanged. This week we will revise the next step on the larger landing page and compare its key-event rate after another complete seven-day window.

Notice what it does not say. It does not claim organic traffic caused the revenue decline. It names the commercial outcome, the associated behavior evidence, the ruled-out refund explanation, and one test.

When to break the routine

Twenty minutes is not enough for every situation. Open a separate investigation when:

• data freshness or connection health is uncertain;
• a payment, currency, or MRR reconciliation does not balance;
• a tracking release changed event definitions;
• the change is concentrated in sensitive customer-level records;
• the proposed conclusion requires identity-level attribution that is not available.

The weekly review should identify these cases, not solve them by guesswork.