
Context: The Brand Holding Two Keyword Exports That Disagree
Picture a Johannesburg online retailer that has exported three spreadsheets in one week. The first is a Google Keyword Planner export showing that a head term commands 8,000-12,000 monthly searches (a wide band, not an exact number). The second is a Search Console performance export showing a completely different set of long-tail queries bringing in clicks it never planned for. The third is a paid search-terms report that agrees with neither, surfacing competitor-intent phrases and misspellings. All three are 'right' and all three are wrong at the same time, because none share a common definition of keyword, intent, or owner.
Every volume number in this article is an estimate or a band. Keyword Planner reports rounded ranges, and third-party tools model rather than measure. Treat all figures as directional inputs, never as budgets.
Why a Keyword Research Layer Sits Under Both SEO and PPC
Most teams treat keyword research as a first-step deliverable: it happens once, gets filed, and both channels quietly drift into their own lists. The discipline that fixes this is a shared keyword layer - one master asset that both the SEO roadmap and the paid account draw from, with a single definition of what each term means for the business.
This is the layer where decisions about revenue, not traffic, actually get made. A query that a paid campaign bids on is a known cost per acquisition. A query ranking organically is a decelerating cost. When the same term appears in both, you need a deliberate owner rather than two teams competing on the same SERP. Prebo Digital's SEO and AI SEO and Google Ads teams both work from one keyword map for exactly this reason - attribution accuracy collapses when two channels each keep a private definition of the same intent.
The Master Sheet: Fields That Make One Asset Serve Two Channels
The master sheet does not need to be elaborate, but it needs to hold the fields that answer 'who funds this term, and what do we expect back?'
| Field | Values | Why it matters |
|---|---|---|
| Term | Normalised exact phrase | Dedupe across all exports |
| Intent stage | TOF / MOF / BOF | Routes channel by default |
| Channel fit | SEO / PPC / both | Prevents internal competition |
| Volume band | Rounded range | Sizing, never forecasting |
| Current organic rank | Position + landing URL | Shows overlap risk |
| Paid CPC / impression share | ZAR estimate + % | Sizes paid test cost |
| Illustrative value | ZAR band | Prioritisation (hypothetical) |
Reading South African Search Language: Abbreviations, Vernacular and Price Modifiers
A keyword list imported from a US template will miss most of the South African demand. Local query language is not decoration, it is a ranking and bidding lever.
- Place abbreviations: Jo'burg, JHB, CPT, Durbs, PTA and Sandton appear alongside the full names, sometimes in one query.
- Language modifiers: English, Afrikaans and isiZulu/isiXhosa service terms each carry distinct intent.
- Mobile-heavy 'near me' phrasing that bypasses neighbourhood names entirely.
- Price modifiers: cheap, affordable, 'cost per month', 'price in ZAR' signal BOF intent even on informational vocabulary.
These variants should be separate rows in the master sheet, not folded into one parent term, because they carry different CPCs and different organic difficulty.
Free Versus Paid Keyword Data Sources for South African Budgets
| Source | Cost | Best use | Limitation |
|---|---|---|---|
| Keyword Planner | Free with Ads account | Volume bands, CPC ranges | Rounded ranges, not exact |
| Search Console | Free | Real queries, 16-month window | Only your own site's demand |
| Google Trends | Free | Seasonality, relative interest | No absolute volumes |
| Third-party (Semrush, Ahrefs) | Paid | Competitor gaps, expansion | Modelled estimates |
Reconciling Keyword Planner Volumes, Search Console Queries and the Paid Search-Terms Report
The reconciliation is mechanical. Normalise every term to lowercase and strip punctuation, then join all three exports on that normalised key. Where Keyword Planner suggests a term but Search Console shows no impressions, the demand may be new or unmet. Where Search Console shows queries with impressions but no Planner history, you have discovered a real local pattern worth investigating. Where the search-terms report shows a paid query converting at a good cost, that term is a strong candidate for an organic content pillar.
normalised_term = term.lower().strip().replace(' ', ' ')join_key = normalised_term # one key, three exportsmaster = planner.merge(search_console, how='outer', on='join_key')master = master.merge(paid_terms, how='outer', on='join_key')master['channel_fit'] = master.apply(route_by_intent, axis=1)Routing Each Keyword Cluster to the Channel That Should Fund It First
Once the master sheet is clean, the routing question becomes: which channel should pay for this cluster now? The default is deliberately simple, because arguments about channel ownership are expensive.
| Cluster type | Intent | First funding channel | Rationale |
|---|---|---|---|
| How-to / definition | TOF | SEO | Low commercial value per click |
| Comparison / alternatives | MOF | SEO first, PPC test | Content plus retargeting audience |
| Head product term | BOF | PPC | Portals and marketplaces contest it |
| Brand + price modifier | BOF | PPC | Immediate, measurable CAC |
Head terms on a South African SERP are frequently owned by portals and marketplaces. That is a signal to buy the term while building organic authority, not to ignore it.
Building a Shared Keyword-to-URL Map So Paid and Organic Land on the Same Page
The most expensive error in shared keyword work is bidding on a thin landing page while a strong organic page ranks elsewhere. A keyword-to-URL map fixes this by naming one canonical destination per cluster and requiring both channels to use it. Pair it with a naming convention so reporting keeps the channels separate.
cluster_id: sa-payments-gateway
canonical_url: /payments/gateway-comparison/
paid_url: same as canonical (no separate PPC page)
content_owner: SEO team
paid_owner: Google Ads team
review_date: quarterlyIf affiliate pages sit on a subdomain or a separate path, exclude them from GA4 conversion counting so organic query data is not diluted by paid attribution. One canonical page per cluster is the simplest guardrail.
Where SEO and PPC Keyword Strategy Conflicts on a South African SERP
Conflict usually surfaces in three places: bidding on terms you already rank first for, cannibalising organic traffic with paid ads on the same query, and creating pages per keyword instead of per intent. The resolution is a documented rule - for example, if organic rank is top three and paid ROAS is below target, pause paid and reallocate; if paid converts but organic rank is beyond page one, keep paid and fund content.
Measuring Both Channels Without Double-Counting: A GA4 View
A term that ranks organically and converts through paid will be counted twice if you read channel reports naively. In GA4, keep the paid search-terms report and the organic query report as separate views, use channel-group breakdowns for blended reporting, and attribute conversions to the last non-direct click. This is the same principle behind Prebo Digital's custom reporting approach: revenue over traffic, accuracy over platform-reported conversions.
A Consent and POPIA Note on Query Data Quality
Consent mode and POPIA compliance reduce the completeness of query-level signal, particularly for users who decline tracking. Budget for modelled conversions and accept that some search-term granularity is lost. Do not overbuild a workaround; a clean consent banner and a documented measurement gap is more defensible than a clever workaround.
Outcomes: What a Working Shared Keyword Asset Looks Like After 90 Days
After one quarter you should see: a single master sheet with one row per normalised term; a documented channel owner for each cluster; a canonical URL per cluster used by both channels; a paid search-terms report that informs the organic content roadmap; and a GA4 view that reconciles rather than double-counts. Prebo Digital, a Johannesburg-based agency founded in May 2016 by an ex-Google employee and a Google Premier Partner (2022-2026), builds this keyword layer as part of strategy consulting. For examples of how this fits together in practice, see the case studies.



