peter/ukaiautomation
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

SEO schema cleanup + blog index update

Browse Source 
Removed 42 deprecated/restricted schema blocks across 21 files:
- FAQPage removed from all commercial pages (restricted Aug 2023)
- HowTo removed from all pages (rich results removed Sep 2023)
- Compliance guide: author type fixed Organization->Person

Blog index:
- New article cards: ai-web-scraping-2026, web-scraping-lead-generation-uk
- Stats updated: 55+ articles -> 57+, 2025 Content -> 2026 Content
- Featured article date updated to March 2026
- Blog schema updated with new BlogPosting entries
This commit is contained in:
Peter Foster
2026-03-08 10:48:11 +00:00
parent 790ffef935
commit 62e69542b0
21 changed files with 40 additions and 867 deletions
Download Patch File Download Diff File Expand all files Collapse all files

82
blog/articles/predictive-analytics-customer-churn.php
View File

@@ -109,46 +109,7 @@ $read_time = 14;
"inLanguage": "en-GB"
}
</script>
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What is the best machine learning model for churn prediction?",
"acceptedAnswer": {
"@type": "Answer",
"text": "For B2B SaaS churn prediction, Random Forest and XGBoost typically perform best, achieving 80-90% accuracy. The best model depends on your data quality and feature engineering. Start with logistic regression as a baseline, then test ensemble methods."
}
},
{
"@type": "Question",
"name": "How far in advance can you predict customer churn?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Most effective churn models predict 30-90 days in advance. 90-day prediction windows give enough time for intervention while maintaining accuracy. Shorter windows (7-14 days) are often too late for effective retention campaigns."
}
},
{
"@type": "Question",
"name": "What data do you need for churn prediction?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Key data includes: usage metrics (login frequency, feature adoption), billing history, support ticket volume, engagement scores, contract details, and customer firmographics. The more behavioral data you have, the more accurate your predictions."
}
},
{
"@type": "Question",
"name": "What is a good churn rate for B2B SaaS?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Annual churn benchmarks for B2B SaaS: 5-7% is excellent, 10% is acceptable, above 15% needs attention. Monthly churn should be under 1%. Enterprise contracts typically see lower churn (3-5%) than SMB (10-15%)."
}
}
]
}
</script>
</head>
<body>
@@ -1766,45 +1727,6 @@ $read_time = 14;
</script>
<script src="../../assets/js/cro-enhancements.js"></script>
<script type="application/ld+json">
{
"@context": "https://schema.org",
"@type": "FAQPage",
"mainEntity": [
{
"@type": "Question",
"name": "What is predictive analytics for customer churn?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Predictive analytics for customer churn uses machine learning models to identify customers who are likely to cancel or stop using your service, allowing you to intervene proactively. Models are trained on historical behaviour data including usage patterns, support tickets, billing history and engagement metrics."
}
},
{
"@type": "Question",
"name": "How much can predictive analytics reduce churn?",
"acceptedAnswer": {
"@type": "Answer",
"text": "UK B2B SaaS companies using predictive churn models typically see churn reductions of 25-35%. Results depend on model accuracy, the quality of intervention strategies, and how early at-risk customers are identified. The 35% figure assumes a well-tuned model with a 90-day prediction horizon and an active customer success programme."
}
},
{
"@type": "Question",
"name": "What data do you need to predict customer churn?",
"acceptedAnswer": {
"@type": "Answer",
"text": "The most predictive features include: product usage frequency and depth, support ticket volume and sentiment, login frequency, feature adoption breadth, billing history and payment failures, NPS scores, and contract renewal dates. You need at least 6-12 months of historical data with known churn outcomes to train a reliable model."
}
},
{
"@type": "Question",
"name": "Is predictive churn modelling suitable for small UK SaaS companies?",
"acceptedAnswer": {
"@type": "Answer",
"text": "Yes, though you need a minimum dataset of roughly 500-1000 historical churn events to train a statistically reliable model. Smaller companies can start with simpler logistic regression models and progress to gradient boosting as data accumulates. Even basic early-warning scoring outperforms reactive customer success approaches."
}
}
]
}
</script>
</body>
</html>