Chapter 29 — Key Takeaways
A one-page reference. If you remember nothing else: report the few numbers that connect search to the business, freeze a baseline to prove change against, and tell the truth — including what didn't work. A report that impresses but doesn't inform, or that hides a bad month, will kill a good program faster than bad SEO ever could.
The one idea
Reporting is where SEO earns its budget, and it is won with fewer, better numbers, not more impressive ones. Pick the five KPIs that tie search to revenue, measure them against a frozen baseline, assemble them into one honest dashboard, tell the traffic → leads → revenue story without overclaiming attribution, and report the failures as plainly as the wins. The blemishes are what make the wins believable.
The five KPIs that matter (and their limits)
| KPI | Source | Why it's a KPI | What it does NOT prove |
|---|---|---|---|
| Organic sessions (esp. non-branded) | GA4 | The fuel; non-branded = new demand captured | A visit is not a customer |
| Organic conversions (calls/forms/bookings/sales) | GA4 | The bridge to money; the number owners feel | Undercounts; last-touch attribution |
| Priority rankings (a small tracked set) | GSC / tracker | Leading indicator; moves first | One query ≠ the market; varies by place/device |
| Indexed pages | GSC | Can't rank a page Google won't index | Quantity, not quality (index bloat) |
| Core Web Vitals (LCP/INP/CLS status) | GSC / CrUX | A fixable page-experience floor | A tiebreaker, not a growth engine |
Before any KPI means anything, freeze a baseline — the starting-point numbers, dated and never edited. Without it, "up 40%" is a feeling, not a fact.
Vanity metrics — keep them OFF the report
| Vanity metric | Why it's hollow | Report instead |
|---|---|---|
| "We rank for 10,000 keywords" | Mostly page-4 terms nobody searches | Priority-term positions + non-branded clicks |
| Third-party Domain Authority / DR | A vendor's prediction, not a Google signal; gameable | Referring domains that send traffic; conversions |
| Raw impressions in isolation | Being seen ≠ being clicked or converting | Clicks + CTR on priority queries |
The test that unmasks any vanity metric: would moving it change a decision? If no, it is decoration.
Leading vs. lagging (the conversation that saves the account)
- Leading indicators move first: impressions, indexing, priority rankings.
- Lagging indicators follow — over quarters: traffic, leads, revenue.
- Rankings up + revenue flat in month three is the expected shape, not a failure. This is the "valley of disappointment" — set the expectation up front, in writing, or the owner cancels one month before the curve turns.
Building the dashboard (free)
- Search Console = search side (impressions, clicks, CTR, position, indexing, CWV).
- GA4 = site side (organic sessions, conversions, engagement).
- Looker Studio = stitch them into one auto-updating screen, KPIs shown three ways (this month / last month / baseline).
- They will never match. GSC counts a click on the SERP; GA4 counts a tagged session on your site (missing bounce-before-tag and consent-declined visits, adding non-Google engines). Pick one source per metric, label it, and explain the gap rather than hiding it.
Cadence
| When | Who | About |
|---|---|---|
| Weekly | You | Monitoring — catch problems (GSC ritual, Ch 27); don't report |
| Monthly | Owner/manager | The KPI dashboard + narrative — the core deliverable |
| Quarterly | Decision-makers | Trend, strategy, budget |
| Annual | Owner/CFO/board | The business case, YoY, the compounding story (Ch 39) |
Same date, same format, every time. Predictability is trust.
Report honestly
- Include what didn't work — the flopped cluster, the flat test.
- Explain the downs you didn't cause — a core update (Ch 6), an AI Overview taking clicks (Ch 36). An explained dip builds credibility; a hidden one is a time bomb.
- Say "organic-attributed," never "organic-caused." Attribution is a model, not proof.
- A report with zero misses is a warning sign, not a triumph.
- Never promise a ranking or a date. Show the trajectory.
Rules of thumb
- Fewer, better numbers. Five KPIs beat thirty metrics.
- No baseline, no proof. Freeze it on day one; never edit it.
- What decision would this change? Ask it of every chart before it reaches the owner.
- Lead with what they feel (leads, dollars); back it with KPIs.
- The blemishes are the credibility. Honesty is what funds year three.
What you can do today
Open GA4 and Search Console and write down last month's organic sessions, organic conversions, clicks, impressions, average position, and indexed-vs-not pages in a dated file — then don't touch it. You have just created a baseline, the thing most programs never make and later wish they had. Next, spend twenty minutes in Looker Studio wiring four of those numbers into one shared, auto-updating screen. Those two habits, built on a calm day, are what let you report the truth quickly on a hard one.
Where this goes next
→ Chapter 30 takes apart the one KPI we kept treating as a single number — "rank." It shows why a universal rank no longer exists (personalization, location, device), how to track positions as a trend not an obsession, how to find your real search competitors, and how to choose tools without wasting money.