IPRATE Methodology

We compare like with like. Ratings are produced separately by country, by practice area (trademarks, designs, patents), by filing track where applicable (national or European), and by time view.

We focus on work done for applicants from the same country as the rated attorney or firm. In simple terms, Lithuanian professionals are mainly judged on Lithuanian client work, German professionals on German client work, and so on.

When a matter has several applicants or several classes, we avoid letting one large filing behave like many unrelated cases. Credit is shared so the ranking reflects sustained practice, not just unusually bulky filings.

Every published ranking can be shown in two time views.

• Long-term: the broad historical view of a professional or firm's track record.
• Emerging: a recent-window view focused on the last five years, designed to highlight current momentum.

These two views answer different questions. Long-term reflects depth and consistency over time. Emerging highlights more recent activity and results.

The core rating is built from six evidence-based pillars. Not every pillar is equally strong in every practice track, but we always use the measures that are meaningfully observable for that market slice.

• Experience: how much real case work is handled.
• Initial success: how often mature cases reach a positive first outcome.
• Relative efficiency: how quickly successful matters move compared with peers in the same market.
• Asset durability: how often granted rights stay alive long enough to show lasting value.
• Renewal or maintenance survival: whether rights are kept in force where that signal is reliably visible.
• Local applicant share: how much of the work comes from the home market rather than mostly foreign inflow.

Some long-tail maintenance signals are thinner in certain patent datasets, so in those cases the rating leans more on the measures that are consistently observable.

We do not compare raw counts from a very large market to raw counts from a small one. Instead, each attorney or firm is compared against peers in the same country, practice area, filing track, and time view.

That means a strong result in a smaller market can still be recognized, even when the absolute number of filings is lower than in a larger jurisdiction.

We also soften very small samples toward the middle. A handful of lucky cases should not instantly create a top ranking, and a handful of unlucky ones should not destroy a solid practice.

Not every name in the source data is published with a rating. We only publish when there is enough evidence to support a meaningful comparison. Attorneys and firms must meet minimum activity thresholds to appear in the public rankings.

That means we check both activity volume and whether enough core measures are applicable for that person or firm. Very sparse or one-off activity is filtered out rather than dressed up as a precise ranking.

Each published result also receives a confidence grade (A through C) based on how much filing data supports the rating. Attorneys with thin data are shrunk toward the cohort median via Bayesian credibility adjustment. Where firm attribution is incomplete, confidence may be capped even if the visible score looks strong.

We reward sustained practice more than isolated bursts. A professional who appears only occasionally over a long period is less likely to be published or to rank strongly than someone with a steadier pattern of work.

For trademarks and designs, long inactive gaps can reduce the final result and, in extreme cases, suppress publication. Patents are handled more carefully because the life cycle is longer and slower.

After combining the applicable measures, we assign one of four public rating labels inside the relevant peer group. Published attorneys and firms appear as Elite, Leading, Proven, or Select within each country-practice cohort.

Rating tiers are always local to the peer group being shown. An Elite result in one country is not meant to imply the same raw volume or market size as an Elite result somewhere else.

Every rated firm, from Select through Elite, has already cleared IPRATE's quality thresholds. The tier ranks firms within an already-filtered pool. Being rated at all means a firm has passed volume, consistency, and confidence filters that exclude the majority of firms in each market.

IPRATE intentionally rates only a fraction of firms in each market. Typically only about 10–20% of firms filing in a country meet our minimum thresholds and appear in the public rankings. The rest drop out at one of the selection stages below.

Common reasons a firm may not be listed:

• The firm filed fewer than 10 cases (cumulative) or 3 cases (emerging) in the country / vertical.
• The firm's filings concentrate in a single class or client, producing thin cross-pillar coverage.
• The firm's data produced only a grade D confidence score.
• Individual attorneys filing in their own name under a solo practice are rated as firms (not separately as attorneys). Solo attorneys may still appear with their firm rating.
• Applicants filing their own IP (self-filers) are excluded. IPRATE rates professional representatives, not in-house filing activity.
• Coverage may be incomplete for countries where national filings are not yet ingested — only European (EUIPO / EPO) filings are covered in those cases.

If you believe your firm should appear but does not, please contact [email protected].

Where direct firm-level case attribution is strong enough, firms are rated from their own attributed work.

In some trademark and design situations, linked attorney evidence can also help build a firm view when direct public attribution is thinner. That keeps useful firms visible while still marking lower-confidence cases appropriately.

Because firm attribution is not equally complete in every dataset, the confidence grade matters especially at firm level.

Claimed profile owners may improve approved presentation details such as descriptive text, website links, and public-facing media.

Those updates can refresh how a profile is presented, but they do not alter the underlying score, rating tier, confidence grade, or methodology.

Any change that would affect the rating itself belongs to the scoring pipeline, not to profile editing.