When does Bulk Generation pay off?
Bulk pays off when every article needs the same structure. Classic use cases: local landing pages ("SEO Agency Berlin", "SEO Agency Hamburg" …), product comparisons, glossary entries, "What is" articles. For high-quality opinion pieces or personality-driven content, Bulk is wrong — you need individual briefs.
Prepare your CSV file
The CSV needs at least a "keyword" column. Optionally add extra fields per row that flow into the prompt — e.g. city, audience, USP, internal links.
keyword,city,audience,internal_link
"seo agentur",berlin,"kmu","/leistungen"
"seo agentur",hamburg,"kmu","/leistungen"
"seo agentur",münchen,"startups","/startups"
"seo agentur",köln,"e-commerce","/shopify"
Tip
UTF-8 encoding and commas as separators are mandatory. Excel often defaults to semicolons and Windows-1252 — check in a text editor before upload.
Upload CSV and check mapping
In the "Bulk Generation" module, click "Upload CSV". Rankion shows the first 5 rows as preview and asks which column is the keyword and which optional variables flow into the prompt.
Define the prompt template
Use {{ double curly braces }} for CSV variables. Example template:
Schreibe einen 1.500-Wort-SEO-Artikel über "{{keyword}} {{city}}".
Zielgruppe: {{audience}}.
Struktur: H1 mit Keyword + Stadt, 4 H2-Abschnitte (Leistungen, Preise,
Vorteile, FAQ), abschließender CTA mit Link auf {{internal_link}}.
Tonalität: professionell, aber nicht steif. Verwende lokale Bezüge.
Tip
Test your template with 3 rows first before generating 100+. A bad template multiplies.
Start and monitor the batch
Click "Start Batch". You land in the job overview with live progress: how many articles are done, in progress, failed. Leave the page anytime — the job continues.
Heads up
Bulk generation costs 8 credits per article. 100 articles = 800 credits. Make sure your budget is sufficient — otherwise the job aborts mid-run.
QA at scale
With 100+ articles, reading each manually is unrealistic. Use the Bulk QA tools.
- › Random sampling — flag random 10% of the batch for manual review
- › Duplicate detection — flag articles with > 30% text overlap (common with too-generic templates)
- › SEO score filter — isolate all articles with score < 70
- › Bulk republish — regenerate flagged articles with adjusted prompt
Result
Result: 100+ consistent articles in an hour, with built-in QA workflow instead of blind trust in AI.
What you learned
- ✓ Clearly distinguished Bulk vs Single use cases
- ✓ Prepared CSV with required "keyword" column and optional template variables
- ✓ Defined prompt template with {{variables}} and tested upfront
- ✓ Calculated credit budget before start (8 × article count)
- ✓ Set up QA workflow with sampling, duplicate detection, and SEO score