Data Limitations – What to Do When the Data You Need Isn’t There


At Antilles Economics we often encounter data limitations both when doing work for clients as well as when conducting our own independent analysis. Not all of these limitations are within our power to fix. But some are. Here’s a selection of common limitations and how we’ve fixed them.


There are times when a particular variable is not explicitly available

Either it is available but not published or it has never been estimated. In some of these cases, there is sufficient data available to estimate it ourselves. This solution may not result in the exact same numbers that would have been published by the authorities, but it usually gives us a good idea of trends. We check for reasonableness to make sure it is in line with our expectations, and focus our use of the information on the direction and magnitude of the changes we observe in the trend, rather than on the exact number.

Another work around we have used for non-existent data is to collect it ourselves. For example, we conducted our Use of Data survey to understand how data was used in corporations because there was no dataset or insights available on this topic for Barbados. In another example, we conducted a survey for a client because there was no data available on the topic they were interested. Companies use surveys all the time to collect their own data on topics such as customer satisfaction, brand loyalty and consumer preferences.

There are times when the data exists but is inaccessible for some reason

For example, many companies have multiple data collection systems depending on the data. There may be one system for financial accounting, another for tracking sales, another for inventory management, another for customer management, another for social media statistics and yet another for production. Many of these systems do not ‘talk’ to each other. So if you want to know the percentage that one particular sales assistant contributed to overall revenue in any given time period, or whether a particular marketing campaign was more effective in one region/branch or another, it is almost impossible though theoretically the data exists. Some companies employ a useful workaround: an overarching business intelligence solution that captures all of the data from all of the various sources into one ‘super database.’ This is the ideal solution. But most companies are ignoring these types of questions because it is too complicated to attempt to answer them and too expensive to purchase a business intelligence solution. We have encountered this problem and solved it in two ways. The first was to use a business intelligence solution that captures data from innumerable sources into one easy to use system. The other was to request the necessary data from all the different sources as we needed it, and combine them manually in a simple software such as Excel. We obviously prefer the first option.

Some databases were not designed to capture information in formats that are useful for analytical purposes.

Another example of inaccessible data can occur depending on database design. For example, a company may have one client that uses three of their services, but the database captures this client as three separate clients because each time the client chooses a new service, a new account must be created. Imagine an analyst wanting to know how many clients are using multiple services or which services tend to be used together, it would almost be impossible with this setup. One solution would be to sample the company’s clients. You could ask a group of relationship managers to analyse their pools of clients and create a simple table with the name of the clients as the rows and the type of service they use as the column headings. Aggregating this table across relationship managers would give you an idea of how many clients use multiple services and which ones. You wouldn’t need to do this for all clients, just a sample, to get a reasonable estimate.


This is just a selection of some of the data challenges we’ve encountered and how we’ve chosen to solve them. We’d love to hear about your experiences, so please leave a comment after this post telling us what challenged you faced, and if you solved it, how.