Main successes
Organizations have made expected and unexpected improvements to their IATI data. While adding the established humanitarian-specific fields to their IATI data was a known improvement, other changes came from ongoing analysis and discussion. For example, to optimize the ingestion of transaction data, a unique reference was discussed and trialed amongst the five publishers. We hope such improvements lead to more useful data for all, rather than just for FTS ingestion.
IATI publishers have had to become data users. Given the various processes behind IATI and FTS data, organizations have had to focus on where differences in their outputs may occur. In turn, getting IATI data ‘into shape’ has highlighted the importance of using it. Organizations cannot ‘just do IATI’ and expect FTS to ingest it. They must understand what they are saying and doing with their data. CBPF, for example, added all donor contributions to their IATI data, and remodeled their allocations to include ‘total grant’ as a commitment transaction. This has only come from considering how their data is to be used by others.
Data ingestion has led to some automation. As we have tested and triangulated IATI data within the FTS system, the longer-term efficiencies become more apparent. We will be able to create an algorithm to spot patterns both within IATI data and other sources. With this, we will have an indication of how IATI ingestion can automate some workflows, whilst freeing up time on the vital manual curation that is often needed.

Challenges faced
IATI data differs between IATI publishers. While a standard is certainly helpful to harmonize format and structure, the actual data differs in terms of scope and detail. The nature of an ‘IATI activity’ varies across—and sometimes within—each of our five initial partners. Hence, it becomes unrealistic to envisage automating IATI completely.
Organizations already express the same data differently. For most organizations, FTS and IATI data is the responsibility of different teams or people. When analyzing these two sources, we have found subtle and significant discrepancies between their composition. One financial flow in FTS could be made up of several IATI transactions. Consensus is needed on what data is relevant.
Humanitarian planning information does not always reflect operational data. We have run into this from several angles with the ‘project code.’ Projects are vital for understanding the intention of humanitarian plans, which are in turn crucial for FTS to chart their progression. But there is often a disconnect between the data on the projects and the data on delivery. Connecting these dots is a challenge, but not the sole domain of our pilot.

Key learnings
> There is no magic button. We cannot expect the availability of machine-readable IATI data to eradicate what came before. In the beginning, that seemed more plausible, since the key ask was to just do IATI. However, we will need to address the new challenges that have emerged, specifically in terms of the data contract between the producer and user.
> Collaboration improves data literacy. The FTS team has a long established set of relationships with a wide variety of humanitarian actors. Introducing a data standard into the mix can be uncomfortable and challenging to regular workflows. We have learned that iteration and discussion between users can lead to a mutual understanding of needs.
> There is a strong appetite for integration. Our pilot publishers have invested time and money to make this work. There is a real demand to avoid the ‘dual reporting’ of data, in line with the goals of the Grand Bargain. Demonstrating good practice will incentivize others to take similar steps and harmonize data outputs.