JD will present "Using Automation and AI to Shut the Door on Fraud in Market Research" at NEXT 2018, April 30-May 1 in New York.

We continue to see research buyers looking for efficiencies on tracking studies to free up funds for new techniques without losing telemetry on market share and trends. This is happening in a context of consolidation (most recently the Research Now/SSI merger) as well as the emergence of newer providers. These trends are reason to review your sample partners and your execution on critical studies. Programmatic sampling and automation can help reduce cost and improve speed while preserving trends and norms.

When Change is a Four-Letter Word
Tracking and normed studies (like advertising and concept tests that measure against norms) are a special breed. They are among the least sexy and most difficult projects in the world of insights. When they are running well, they seem deceptively simple and thus ripe for scaling back for savings. When they aren't running well, businesses writhe in pain.

The First Commandment of tracking studies is Thou Shalt Not Change. Yet the instability in the broader sample ecosystem, from cratering participation to supplier consolidation to “respondent-dumping”, has made clients audibly (and justifiably) concerned about the reliability of their trackers and normed studies. Any change virtually guarantees the breakdown of trends and norms that serve as performance benchmarks. Researchers have long believed that working with the same sample companies mitigates this problem, but even this is no longer certain. Worse, these studies typically involve repetitive manual operations that make it difficult to improve speed and cost without compromising quality.

Automation Yields Consistency and Efficiency
Automation solves these problems in four important ways.

First, replacing error-prone human labor by faster, set-it-and-forget-it, never-on-vacation machines unlocks huge efficiency gains, reduces costs and improves speed.

Second, by removing their operating duties, companies can redeploy their human experts for insights discovery and value creation.

Third, the consistency that automation brings to research execution usually improves data stability significantly. Automation enables fine-grained control of operational processes to more effectively manage sample blends, spot problems, and correct them in field before they cause headaches.

Fourth, automation can also yield improved response quality, but only if the supplier is focused on this. Not all automation is created equal. Most API solutions do little more than optimize the supplier’s economics. So much more is possible though. When fully built out, automation reduces fraud, improves the respondent experience, and yields better data.

Go Programmatic
Adopting sample automation is easier said than done. While decisions to move forward can ultimately have a positive impact on accuracy and speed, these decisions will impact the research process from beginning to end. Every company has different challenges and goals, so the path to programmatic is not always an obvious one.

It’s vital that companies making this shift consider every angle. Some common issues that need to be addressed head-on when implementing an automated approach include:

  • Taking the time to fully map and maintain the data fields so the systems communicate fluently. This part of the puzzle is time consuming and it’s tempting to take shortcuts. Don’t: you will find that you start to have “leaks” in the system that compromise feasibility. Find a provider that has a lot of experience on this, especially with advanced and custom APIs.
  • Respecting respondents to get quality data. Automation used poorly can bounce respondents around from router to router, resulting in frustration and dropouts. Automation used correctly can actually help improve respondent experience from the ground up.
  • Expand the scope of automation so that all functions, including feasibility and monitoring, are machine driven. Leaving these tasks in the hands of humans as a manual process opens companies up to problems in field that will not be fixed in a timely manner.

The approach to programmatic can be complicated but the right platform (and tapping into the right expertise) will enable all the benefits of automation, while not forcing you to change sample sources.

When implemented to its fullest capacity, a programmatic approach is flexible and dependable. It maintains the integrity of your data by providing greater reliability, efficiency, and error reduction. Advanced APIs can ensure data and suppliers remain consistent and mitigate changes in benchmarks to preserve validity over time. The bottom line? Take a thoughtful, carefully planned approach to programmatic sampling and you can get better quality data, faster and cost effectively with your tracker.