Understanding and checking your SCHUFA score

SCHUFA divides the new score into four bands: 776–999 excellent, 709–775 good, 642–708 acceptable and 100–641 adequate. Two points about this that are rarely mentioned. First: common rules of thumb such as “900 and above is very good” or “below 700 becomes problematic” correspond to none of these bands – anyone who relies on them is relying on thresholds that do not exist. Second, and more importantly: the band says nothing about how a particular company decides. A contract is not decided by SCHUFA but by the bank, the landlord or the provider – according to their own criteria, which include income, employment and internal policies. A “good” score is a necessary, not a sufficient, condition.

Nobody outside SCHUFA knows that, and you should distrust any source that claims otherwise. Since 17 March 2026, SCHUFA has named twelve criteria and states a maximum value for each criterion – that is considerably more transparency than the old model with well over a hundred characteristics. However, the weighting and the formula are still not disclosed. Legally, that is exactly the point of contention: Article 15(1)(h) GDPR (DSGVO) entitles you to “meaningful information about the logic involved”, and whether a generally understandable explanation is sufficient for this, or whether the information must allow for recalculation, was heard by the Federal Court of Justice on 18 June 2026 (I ZR 226/25 to I ZR 230/25) – with no decision so far.

No, and that is the most important practical limitation. The changeover runs until 2028 in stages: at first, only some of the enquiring companies used the new value, the proportion grows step by step, and only after that is the old score fully replaced. Until then, each company decides for itself which value it queries. For you, this means: you cannot know and cannot influence which score was used for a specific enquiry – and therefore cannot say for certain why a decision turned out the way it did. Anyone who tells you “your score” had this or that effect at a particular bank is speculating. What you can do concerns the data, not the figure.

Answered honestly: only what addresses the underlying data reliably works – everything else is guesswork about an unknown weighting. Three things are demonstrably worthwhile. First: have unlawful or incorrect entries removed. Check every negative entry against the standard set by § 31 Abs. 2 BDSG; what should never have been listed there must go. Second: insist on the conditions enquiry when comparing loans, which stays invisible to others. Third: time – deletion periods run out, and nothing can speed that up. What you should not do: replay points tables from the internet or close accounts for tactical reasons. Nobody knows whether or how that works.

Because so-called positive data (Positivdaten) may be transmitted – that is, data that “contains no negative payment experience or other conduct not in accordance with the contract”, such as identification data and information about the conclusion or termination of a contract. The Federal Court of Justice confirmed this on 14 October 2025 (VI ZR 431/24): the transmission is lawful even without consent, because it is justified under Article 6 GDPR (DSGVO) by the interest in fraud prevention, which outweighs the interest in non-disclosure. You therefore cannot challenge such entries on the grounds that you never consented. The position is different for negative characteristics (Negativmerkmale) – there, the strict standard of § 31 Abs. 2 BDSG applies.

Yes, but subject to conditions – and one of them is rarely checked. § 31 Abs. 1 BDSG requires, among other things, that the data used is demonstrably relevant “on the basis of a scientifically recognised mathematical-statistical procedure”, and that not exclusively address data was used for the calculation; where address data is also used, the data subject must be informed beforehand. There is also a European law dimension: the Court of Justice of the European Union ruled on 7 December 2023 (C-634/21) that creating the probability value is an automated decision under Article 22 GDPR (DSGVO) where the value decisively determines a third party's decision to enter into a contract – with the rights under paragraph 3.

Contact the company that refused you first, not only SCHUFA: that is where the decision was made. If it is decisively based on an automatically generated probability value, the safeguards of Article 22(3) GDPR (DSGVO) apply – you can demand human intervention, express your own point of view and contest the decision. The Court of Justice of the European Union expressly clarified in C-634/21 that it does not matter that the refusal formally comes from the company rather than from the credit reporting agency. At the same time, request the free data copy and check whether it is based on an entry that should not have been listed there.

Not immediately, and you should plan for that rather than checking every day. If an entry is deleted or corrected, the change must first reach the database and feed into the next calculation; how quickly that is reflected in the value depends on the weighting, which is not public – so that, too, cannot be predicted. Deletion periods run in months in any case: under the Federal Court of Justice's case law of 18 December 2025, three years apply to settled payment defaults, and 18 months under three additional conditions. Set yourself a reminder for a few months ahead, then check again with the free data copy – and do not make any decisions based on guesswork in the meantime.