Anomalo

Figure 2 – Rich, visual alerts optimised for data quality insights

As well as what it does, which is outlined above, it is important to understand what it does not. Anomalo is aiming at machine-operated data pipelines, so does not have functionality to do traditional human-entered data deduplication, customer name and address validation and enrichment, or merge/matching of duplicate regards. It leaves this kind of thing to more traditional data quality and master data management tools. In the future the company is likely to explore additional functionality such as automatic detection of personally identifiable data, and also monitoring the data quality of unstructured as well as structured data. Interest in unstructured data is a hot topic at the moment as enterprises explore the training of large language models on their own data in order to deploy artificial intelligence projects. In many cases the data that will need to be trained is unstructured, and such data has just as many issues as structured data in terms of its quality. AIs are sensitive to the quality of the data that they are trained on, so any product that can significantly help in this area is likely to have a bright future.

Certain types of data situations are hard to deal with if you rely on manually entered data quality rules. One Anomalo customer has five million tables that it needs to monitor, and estimated that it would take them over 1,700 person-years of effort to manually code all the data quality rules that it would need. In such large-scale cases, some form of automated approach is the only practical way forward.

Anomalo competes directly with Monte Carlo and with more traditional data quality software suites from vendors like Informatica and Ataccama. It focuses primarily on data that is machine-captured and high in volume, where manual data quality rule creation is impractical to scale. Traditional data quality vendors tend to focus on areas like customer name and address validation or product information management, where data like a new customer is typically entered by a human being, with all the potential risks for error and duplication that implies. Anomalo is particularly suited to situations where the number of records in the data pipeline is huge, and where automation is the most practical way of detecting data issues rather than manually coding rules. One product differentiator is the ability of the product to not just monitor data pipeline flows but also dig into the data itself to highlight unusual data values, something that not all its competitors can do at present.

The bottom line

Anomalo is a fast-growing vendor in the emerging data observability space within the broader data quality market. If you have a high volume of machine-collected data to deal with and monitor, then Anomalo should be on your shortlist of vendors to consider.