Data collection is not always treated the way it should. A common misbelief is that it is quite simple: Get the sensor, get the hardware, connect them and done. But it isn’t that easy.

Refinements to the end product are based on collected data, meaning incorrect measurement practices can curb efforts into the wrong venues and waste resources on problems that may not even exist.

There is no golden rule for an immaculate data collection approach, but there are simple strategies that guarantee success.

Below are five tips for getting started on your data collection journey:

1. Know What You Need and Why You Need It

Before collecting data, know what kind of data you need and why. At the end of the day, data collection is going to serve a higher purpose, whereas it is product development, health monitoring or safety. The first step is to figure out how the collected data is going to be used. This will give you an idea of the types of physical quantities that need to be measured and what sensors are required, as well as how fast and accurately you need to measure the data.

2. Do You Really Need It?

Data collection hardware can be extremely expensive. Some aspects of data collection hardware like reliability, repeatability, reproducibility and accuracy are non-negotiable: Hardware incapable of satisfying these preliminary requirements must be avoided at all costs. The good news is that you might not need that shiny high-end system for your application after all. Lab-grade sensors for a shop floor data collection system are often unnecessary.

3. Your Setup Matters

Sensor placement is at the root of most measurement inaccuracies. For instance, when trying to measure fluid temperature in a pipe, you must consider the time lag between the fluid temperature change and when the sensor registers it. The fluid has to heat up or cool down the pipe and, depending on thickness and airflow over the pipe, you can get delayed or inaccurate readings. Although sensors are wonderful devices, they are not smart – measuring is all they do.

The way signals get to data collection systems also matters, mostly because noise can be an issue. Signal noise can be a result of many factors, including loose connectors and electromagnetic emissions. If the signal level is not that high, it can become a challenge. Fortunately, routing measurement cables separately from control or power cables limits these errors.

4. Know Your Data

Trust your data. If your data does not make sense, there is a good chance something is wrong. More often than not, the issue is either with the system you are collecting the data from or the data collection system itself. But do not be fooled. Avoid interpretations that make insensible data look normal or justified. Confirmation bias is your number one enemy, so always play devil’s advocate.

5. Keep Your Data Close, but Your Data Collection System Closer

Even when everything is going your way, there is still a lot of work to be done. Performing maintenance on your data collection system is crucial, including periodic inspections of cables, sensors and connections and the calibration of sensors and the system.