A ranking is a diagnostic. Revenue is the scoreboard.

A finance director asked us exactly that on a review call last year, and the SEO lead in the room did not have a good answer. Position four sounds like progress. It says nothing about whether the business made more money because of it, and finance directors are paid to notice that gap.

The formula that survives that kind of scrutiny is: SEO ROI = (value from organic search − SEO investment) ÷ SEO investment × 100. Value means attributable ecommerce revenue, collected revenue from CRM opportunities, or a deliberately conservative estimate of qualified-lead value while better data is still arriving. A ranking can explain why visibility changed. On its own it is not value, and it should never be presented as if it were.

That definition needs a fixed window to mean anything. A monthly report suits operating decisions. A rolling three- or six-month view suits judging a programme that compounds unevenly, which almost all of them do. New domains and competitive South African service categories rarely produce a tidy month-one return, and that is not a reason to skip measurement, it is a reason to date your baselines and assumptions carefully so nobody re-litigates them from memory later.

Count the full investment honestly: strategy, technical work, content production, outreach where relevant, analytics implementation, and agency or team time. Reporting only the retainer line while treating the developer’s hours and the copywriter’s time as free produces a flattering number that will not survive a second look from anyone who actually does the maths.

Every business collects softer signals — scroll depth, newsletter sign-ups, PDF downloads, engaged sessions. Keep them for diagnosis, but keep them separate from the event that actually signals sales potential. A B2B consultancy might choose a completed discovery form; a clinic, a booked appointment; an ecommerce store, a completed purchase. A WhatsApp conversation only counts as a primary conversion when a real person answers it and logs the outcome.

Where qualification happens after the click, use two stages: an organic lead, and an organic qualified lead. This stops marketing celebrating raw form volume while sales quietly rejects most of it. Agree on the definition together — budget, service fit, location, decision-maker status or appointment attendance can all matter. The right definition predicts a real opportunity. It does not simply make the dashboard look busier than it actually is.

Configure a stable event for every action that matters, confirm it fires in DebugView, and mark the commercial ones as key events. Avoid event names that shift every time a campaign or a page redesign happens, a durable naming convention is what makes a year of reporting comparable at all, and renaming events every quarter quietly breaks every trend line you were relying on.

For a South African lead-generation site, track at minimum a lead-submitted event on a successful form, a whatsapp-click event on a tap to a prefilled WhatsApp link, a phone-click event on tap-to-call, and a booking-confirmed event where a booking exists. Store useful context — landing page, service interest, form type — but keep personal data out of GA4 parameters entirely.

Connect Search Console to GA4 for a shared view of search performance, then use Search Console on its own for query and page opportunity. GA4 tells you what a visitor did after arriving. Search Console tells you where Google is already giving you impressions or clicks. Neither replaces the CRM, which is where quality and revenue actually get confirmed by a human being.

• Test each conversion on desktop and mobile, including the thank-you page or the server-side success response.
• Exclude internal traffic, and define referral exclusions for payment, booking or form tools that can quietly overwrite the session source.
• Apply UTMs consistently to paid and email activity so organic is not credited for traffic that was never actually organic.
• Note consent-mode and cookie-banner gaps in the report. Missing data is a caveat you disclose, not a reason to invent precision that is not there.

Keep the executive scorecard short. It should show inputs, outputs and the economic consequence — nothing more. The delivery team can hold a deeper working dashboard; leadership needs the handful of numbers that support a scale, hold or fix decision and nothing they will have to ask you to explain twice.

Segment brand and non-brand demand where you can. A surge in people searching your company name may be a healthy side-effect of other marketing, but it should not be credited to new SEO content that had nothing to do with it. Non-brand performance is usually the clearer read on category-demand capture. For location-led firms, also segment by service area and track calls, direction requests and website clicks from Google Business Profile alongside the web conversions.

Last-click attribution gives SEO clear accountability when the organic visit produced the final tracked conversion. It also under-credits early research content that started the journey. First-click has the opposite bias. Data-driven attribution can help once enough clean conversion volume exists, but it is never a substitute for a CRM that actually records source and outcome by hand.

A practical approach reports two numbers side by side. First, last-click organic conversions and revenue. Second, the number or value of opportunities where organic assisted along the way. Keep the two methods visibly separate, so leadership can see a direct return without pretending a buyer who read three articles, clicked a LinkedIn post, and later came back through branded search had one simple, single-touch source.

Do not credit every deal that ever visited an organic page to SEO. Define an attribution window that reflects your actual sales cycle, document it, and keep it consistent. For high-consideration financial, legal or professional services, an assisted-conversion view usually tells you more than trying to force every journey into last click.

Illustrative scenario, not a client result: a Johannesburg B2B services firm spends R18,000 in a month on SEO work, content and implementation included. GA4 records 36 organic form or WhatsApp leads. Sales qualifies 18 of them. Their historical close rate from qualified lead to sale is 25%, and average first-year gross profit per new client is R12,000.

The model is not claiming 4.5 people literally became clients. It estimates expected value from observed lead quality. As closed-won data arrives, replace the assumed close rate and value with your actuals. Use gross profit or contribution rather than top-line revenue where margins vary materially — revenue alone can make a genuinely weak channel look profitable on paper, which helps nobody once the next budget conversation starts.

Organic traffic can rise while ROI falls, if the page answering the query does not actually help the visitor take the next step. Review organic landing pages by sessions, key-event rate and qualified-lead rate. A high-impression page with a low click-through may just need a sharper title and meta description. A high-click page with low conversion may have an offer, proof, speed or CTA problem sitting underneath it that no amount of extra traffic will fix.

Match the action to the intent behind the search. Someone searching "SEO agency Johannesburg pricing" is closer to a consultation than someone searching "what is technical SEO." The first page can earn a comparison, proof and an enquiry CTA. The second should give a genuinely useful answer and a relevant next step, without forcing a sales pitch onto someone still researching. Internal links help both a user and a search engine move from research to decision.

This is exactly why content, UX and technical SEO should never be reported as three isolated workstreams. A technically flawless article reaching the wrong audience has weak value on its own. A persuasive service page that cannot be indexed has none at all.

Monthly discipline prevents the two failures we see most often: publishing without learning, and reporting without acting on what the report says. Keep a short decision log — what changed, why, the expected impact, and what the next review actually showed. That log protects institutional knowledge when staff or suppliers change, which they will, usually at the least convenient moment.

1. Week 1 — validate the dataCheck key events, CRM source capture, indexing changes, crawl issues and any tracking break a recent release introduced.
2. Week 2 — diagnose landing pagesReview organic conversion and qualified-lead rate by landing page. Fix the biggest leak before you add more traffic on top of it.
3. Week 3 — act on demand evidenceUse Search Console queries and pages with impressions to sharpen intent fit, internal links, headings and useful depth.
4. Week 4 — make the investment decisionShare ROI, assisted context, work completed, risks and one clear next action: scale, continue, repair or pause.

Quarterly, revisit your lead-value assumptions, compare organic cost per qualified lead against paid alternatives, and assess whether the business can actually serve the demand this category represents. SEO should earn a bigger budget only when both the company’s capacity and the measurement support that conclusion together.

Vanity reporting is usually a cover for an incomplete measurement system. Fix the system instead of dressing up the report. Do not use average keyword position as a business KPI — it mixes queries, locations and devices, and says nothing about intent or value. Do not report "traffic up" without showing landing-page quality alongside it. Do not call every enquiry organic when the CRM source field is simply blank because nobody filled it in.

• A named owner checks conversion tracking after every website release.
• Sales records lead status, opportunity value and closed outcome in a shared system.
• The monthly report states both cost and the attribution method used.
• Organic results are segmented by non-brand, brand and local intent where relevant.
• Assumptions carry a date, stay conservative, and get replaced with actuals as they arrive.
• Every dashboard review ends with a named owner and a deadline for the next improvement.

The goal is not to make SEO look good. The goal is to know whether organic search is creating profitable demand — and what has to change when it is not.

Thandiwe Nkosi, SEO & Content Strategy Lead

If tracking is missing, start with instrumentation and a baseline month before anything else. If traffic exists but leads do not, audit offer clarity, CTA paths and mobile experience before commissioning another batch of content. If qualified organic leads are appearing but close rates are weak, sit marketing and sales down together and check expectations, pricing, follow-up speed and lead qualification honestly, without either side getting defensive about it.

For a local service business, make Google Business Profile, review requests, service-area relevance and call handling part of the same plan. For a national B2B firm, weight non-brand commercial pages, authority content and CRM lifecycle attribution more heavily. The formula stays the same; the implementation follows how your customers actually buy.

A trustworthy SEO ROI report should make the next decision easier, not harder to have. When it does not, the fix is usually cleaner definitions and better instrumentation — not a more colourful chart.