What are Transfer Pricing Data Points, And 5 reasons to have them structured
Data in transfer pricing
In digital transfer pricing, data is essential. Data is different types of information that usually is formatted in a particular manner. For example, data can be in the form of words, numbers, images and so forth. And generally, there are two types of data: unstructured and structured data.
Unstructured data
Unstructured data is where the information is available in a wide range of formats, and a computer cannot quickly process it. It may need to be processed and stored by humans, who can analyze the information. Here are examples of unstructured data from a transfer pricing world:
Example 1: Extract from the OECD Guidelines 2022, p.333
Why is it unstructured data? Because it is raw text. To make things worse, it’s not even a text, but a picture of the text. It has certain logic and facts embedded, but it is not machine-readable – you can’t give a computer task to list allocation keys based on it (at least without using advanced AI technologies).
Example 2: Description of the transfer pricing case
Same as above, this is raw text with a certain logic, but only experienced humans can read and make sense of it.
Structured data
Structured data, on the other hand, is where data is usually structured in a particular format which makes it easier to analyze or process. It is available in a specific format like a spreadsheet or database, which computer programs can easily read or extract information from. Let’s modify one of the previous examples to make it structured:
Now, the transfer pricing scenario from Example 2 is presented in a structured way – the information has specific characteristics (dimensions) and is formatted in the table. Each separate cell in the “data” column is a data point (BOLD). And since it is structured, we call it a structured data point.
Why do you need structured transfer pricing data points?
Traditionally, unstructured data is used in transfer pricing, which makes the entire process manual and error-prone. But the critical issue is that even if the information is available in the tables, it is not structured so that computer programs can easily analyze and make sense of it.
The answer lies in using structured data as part of a digital approach to transfer pricing. If a company collects structured information and stores it in a computer system, transfer pricing or financial specialists can easily analyze it and use it for transfer pricing policy and compliance purposes. Here are a few benefits of having structured data in transfer pricing:
Structured data is easy to search and analyze. For example, you may want to know the sum of your intercompany transactions using a transactional net margin method - and you can get this data in a few clicks. In the case of unstructured data, though, you may need to go and open dozens of TP reports, read them and copy-paste specific values. In other words, you get unique insights.
Structured data can be easily used for various applications. For example, you can generate many TP local files in a few clicks if the computer knows which data to put in a template. Another use case is preparation of TP forms and digital local files (like in Australia, Poland or Belgium) - they are all in XML format which is ideal for structured data.
Structured data allows updating in real-time. For example, when you update the TP files, you can insert the data automatically into all of your documents.
Structured data is easier to store and reuse. Every transfer pricing specialist knows the pain of having to read the historical local files, often physical copies stored in remote office locations. Having structured data, you can easily extract a summary or the fact you need to know without a need to read hundreds of pages.
Structured data allows for better tracking and workflow management - the whole point of digital transfer pricing is that it helps to improve control and compliance. It's much easier to know what data you need and what's missing if it is structured well
BONUS: how you can structure unstructured data automatically (using cutting edge AI technologies)
In this video, we are showing how GPT-3 AI engine converts unstructured data in Example 1 into a structured list (table) of allocation keys by providing AI with one example only. Imagine how you apply this technology to your raw data! Migration into the digital world was never easier.
About the author:
Borys Ulanenko is a Digital Transfer Pricing Category Lead at Aibidia. Borys has more than 9 years of experience in transfer pricing with a background in industry and consulting. In addition, Borys is the founder of the educational platform StarTax Education. At Aibidia, he focuses on developing new transfer pricing applications and contributes to marketing and business development.
