A metric that evaluates the quality of search results by measuring the ranking of relevant items, giving higher weight to more relevant results appearing earlier in the list.
SEO professionals or data analysts reading about search engine optimization metrics like Click-Through Rate (CTR).
01What it is and how it works
NDCG calculates the effectiveness of a search algorithm by combining two factors: the relevance of results and their position in the ranking. It starts with a Cumulative Gain (DCG) score, which sums relevance scores of results but discounts lower-ranked items using a logarithmic scale. For example, the first result gets full weight, while the second is divided by log2(2), the third by log2(3), and so on. To normalize, this DCG is divided by the ideal DCG (the best possible ranking for those results), producing a score between 0 and 1. A score of 1 means perfect ranking, while 0 indicates no relevant results or poor placement.
NDCG measures how well a search engine ranks relevant results, with better rankings getting higher scores.
02What to do about it
To improve your NDCG score, focus on optimizing content relevance and user experience. First, ensure your pages directly answer common search queries by using clear, concise language and structured data like schema markup. Second, prioritize user engagement metrics (e.g., bounce rate, time on page) as these signal relevance to search algorithms. Third, audit your site’s internal linking to guide users to related relevant content. Finally, monitor NDCG trends in analytics tools like Google Search Console or third-party SEO platforms to identify underperforming pages and prioritize fixes.
03How it is measured or noticed
NDCG is typically measured through search engine analytics tools or third-party SEO platforms. For example, Google Search Console provides metrics like 'Average Position' and 'Click-Through Rate,' which indirectly reflect ranking quality. Advanced tools like SEMrush or Ahrefs may display NDCG-like scores based on relevance and position. To calculate it manually, assign relevance scores (e.g., 0 for irrelevant, 3 for highly relevant) to search results, compute DCG, then normalize against the ideal DCG. A score above 0.8 is generally considered strong, while below 0.5 suggests poor ranking alignment with user intent.
04Common mistakes
- Ignoring position weighting: Assuming all relevant results are equally valuable regardless of rank.
- Overlooking normalization: Comparing raw DCG scores across different query sets without adjusting for ideal rankings.
- Misinterpreting scores: Confusing high NDCG with high traffic, when it only reflects ranking quality.
- Neglecting user intent: Focusing on keyword density over content that genuinely answers queries.
05Limits
NDCG is most useful for evaluating ranking quality in search engines, not for measuring traffic or conversions. It assumes relevance can be quantified, which may not hold for ambiguous or subjective queries. It is often confused with Click-Through Rate (CTR), which measures user behavior, while NDCG focuses on algorithmic ranking. Additionally, NDCG requires a predefined set of relevance scores, which can be subjective or inconsistent across evaluators.
06Worked example
For a query with three results rated 3, 2, and 1, DCG = 3 + 2/log2(2) + 1/log2(3) ≈ 3 + 2 + 0.63 = 5.63. If the ideal ranking (3, 2, 1) has DCG = 5.63, NDCG = 1. If results are ranked 1, 3, 2, DCG = 1 + 3/log2(2) + 2/log2(3) ≈ 1 + 3 + 1.26 = 5.26, so NDCG ≈ 5.26/5.63 ≈ 0.93.
Frequently asked questions
How does NDCG differ from metrics like CTR or MRR?
NDCG focuses on ranking quality by weighting relevance based on position, while CTR measures click-through rates and MRR tracks the highest-ranked result's relevance. NDCG penalizes lower-ranked relevant items more heavily.
Should I prioritize NDCG over other metrics when optimizing search results?
Yes, if ranking quality is your primary goal. NDCG directly evaluates how well relevant items are positioned, making it more relevant than traffic or conversion metrics for this purpose.
How exactly is NDCG calculated?
NDCG uses a logarithmic scale to assign higher scores to relevant items at top positions. It compares the actual ranking to an ideal ranking where all relevant items are at the top, normalizing scores across different result sets.
Does NDCG still work well with AI-generated search results?
Yes, but it depends on the AI's ability to rank relevance accurately. If the algorithm misjudges relevance, NDCG will reflect poor ranking quality, so ensuring high-quality input data is critical.
What happens if my NDCG score is consistently low?
A low NDCG indicates poor relevance or poor positioning of relevant items. You should audit your content relevance, ranking algorithms, or user feedback to identify and fix gaps in how results are presented.
How quickly can I improve my NDCG score?
Improvements depend on changes made. Optimizing content relevance or user experience can show results in weeks, but structural algorithm changes may take longer to impact rankings.
Asked out loud
spoken, not typedThe same term in the words somebody uses speaking to an assistant rather than typing into a box — written from the situation, which is why each one carries the situation it came from.
You can check it through your analytics dashboard or SEO tool. Just log in and look for the NDCG metric under search performance reports.
It’s a ranking quality metric. If your search results aren’t showing the most relevant items first, your NDCG will be low, which means users might not find what they need.
Sure. Open the report, find the NDCG section, and look for trends. A dropping score means your ranking quality is worsening, often due to irrelevant results appearing higher.