Five common data collection challenges
Collecting data is often harder than people expect. Customer surveys, field work and collecting data from different systems can all present problems. Here are 5 common data collection problems and some solutions to get you started.
1. Too much data to manage
Large quantities of data can quickly become overwhelming. Even though you have lots of information available to you, in reality this can create organisational problems, for example where the data is stored, who can find it and how long it takes to analyse it.
When processing large amounts of data, it is necessary to have the right IT infrastructure in place. This could be suitable software and storage systems in your own company or external support from a specialist data collection company. Without it, analysis becomes slow and people lose trust in the results. Prior to collecting data, define the questions that you want to answer with it and only collect the data that is necessary to answer these questions.
2. Data bias
Data bias occurs when the information collected does not accurately represent the group or the situation that is being studied. This can happen for a number of reasons, including how that data is gathered, who is included, and how the actual questions are framed.
An email-only survey misses people who do not regularly use email. Leading questions can be worded in such a way as to elicit a particular answer. Data collected only on weekdays may not capture information from weekend customers. In order to try and reduce bias, consider who else may be excluded from the data, test the questions on a small group first and compare the sample to what you know about the wider population.
3. Lack of quality assurance processes
Without proper checks in place, even a data collection company can encounter errors and inconsistencies.
Error detection during quality assurance is designed to identify errors such as duplicate entries, wrong formats, missing data fields and illogical values. While limited error detection allows small errors to accumulate without being detected and undermine the whole data set, many of them can be prevented with simple steps such as validation rules on input forms or clearly instructing data collectors in the field. Occasional checks on a sample of entries as well as regular cleansing of duplicates also help.
4. Limited access to the right data
A different problem altogether is too little of the right data. That information may be held in another department’s system, by a third party, cost too much to purchase, or simply not exist yet. As a result, you cannot form a complete picture of the subject at hand, and you run the risk of forming a conclusion from half the story and being too confident about it.
List what you need and where this information already exists. It may be possible to share existing information with a suitable agreement in place. Information that does not currently exist may be collected through a targeted information collection exercise such as a short survey or site visits.
5. Legal and compliance requirements
Data collection is governed by legal and regulatory requirements that set limits on how information can be gathered, stored and used. In the UK, personal data is covered by UK GDPR and the Data Protection Act 2018. The regulator, the Information Commissioner’s Office, sets out the data protection principles, including collecting only what you need, using it for a clear purpose, keeping it accurate and secure, and not keeping it longer than necessary.
Make these considerations part of your process from the outset and make clear to others why you are collecting their data and what you will do with it. Keep access to data to as few people as necessary.
The five challenges at a glance
| Challenge | Warning sign | First step |
|---|---|---|
| Too much data | Reports take days, nobody uses them | Collect only what answers your questions |
| Bias | Results do not match what you see in practice | Check who is missing from your sample |
| Poor quality | Duplicates, blanks, odd values | Add validation and spot checks |
| Limited access | Key questions cannot be answered | Map where the data you need lives |
| Compliance | Unclear why data is held or for how long | Set a purpose and retention period |
Frequently asked questions
How much data should we collect?
Collect no more data than is necessary to arrive at reliable answers to your questions. More data is not always better: storage, protection and cleansing all cost more. For personal data in particular, collecting more than is needed is contrary to data protection principles. Collecting more is also likely to result in a less useful data set than one that is small, well-designed and accurate.
Should we use a data collection company or do it ourselves?
Simple surveys can easily be done in-house depending on the size, skills and location of your organisation. More complex research projects, the collection of data at multiple locations and the need for trained field staff typically benefit from utilising a specialist research agency that can supply the necessary staff, equipment and quality control measures. Remember that either way, you are responsible for the data and how it is used.
Where to go next
If your data is stored online, our guide on preventing network attacks covers the basics of keeping it secure.
