At Focaldata AI, we’ve been using our qual AI platform to replace focus groups at three key junctures in the campaign lifecycle.

These are

• ‍Pre-quant questionnaire scoping 🗒️
• Audience deep-dives 🔍,and
• ‍Message testing 🎯

Along the way, we’ve found that our new qual platform goes a long way towards mitigating several known pitfalls of qualitative research, like the long lead times from participant recruitment to insight and the difficulty of extracting deep insights from qual data.

Let’s dive in.

The traditional campaign research flow 

Qual has always been an important part of the political campaigns toolkit.

We’ve been asking people ‘why’ they intend to vote or act in a certain way, through qualitative interviews, since at least the 1940s. 

Before the arrival of AI-powered research platforms, we used focus groups in three ways. 

‍

The first main use case is pre-survey scoping. Not sure what answer options to include in your long multi-select question? Want to test the understanding of key or contested terms before you commit them to a large, expensive sample? You run your survey ideas through a focus group to check your questions won’t land awkwardly. Good survey designers might run 2-3 focus groups among their key audiences before deciding how to field their questions. 

‍

Once the survey is complete, campaign teams will typically use this data to create an audience segmentation. But without qualitative insights, it’s difficult for a segmentation to come to life. Focus groups are used to develop these real-life personas. Qualitative discussions help to ‘fill in’ the fictional representation of these segments – uncovering their most deep-set behaviours and views. This makes it far easier to develop compelling messages and content for them. 

‍

The final application for focus groups in the campaigns toolkit is message testing. We’ve all sat in dozens of focus groups with moderators rotating through different messages, slogans and posters to figure out which one to pick. Some of the best political slogans (“Take back control”, “Get Brexit Done”) have emerged in focus groups settings. Sometimes, a conversational forum is the only way to discover new things (rather than quantifying what’s already known).       

‍

So… what’s the problem?

‍

Simple: focus groups are slow and expensive. The wider qualitative process doesn’t scale. The advantages of qual – depth, richness – are Janus-faced because they are so people and time-intensive to extract. As the diagram above shows, a (relatively fast) data collection cycle for a campaign takes 9 weeks with 6 weeks reserved just for focus groups. By the end, you’ve only spoken to 30 people in-depth. What if your winning slogan came from the 31st person? That could be the difference between a win or a defeat (recent history suggests victories are won at the margin, after all!) 

‍

Enter Focaldata AI: Qualitative insight at the speed / scale of quant

‍

Despite the pace of present-day campaigns, we don’t think you should have to compromise between qual and quant. Each method teases out different dimensions of insight – and we really need both. 

Focaldata’s depth qual AI solves this problem. It provides rich qualitative insights at the speed of quant. This means it’s now possible to have 100s of deep, focus-group like conversations simultaneously - at a fraction of the time and cost of traditional focus groups. 

Using the platform, you can set up a 25-minute discussion with an individual from your target audience groups and have results back within the hour. You can directly query the dataset, ask for verbatim quotes, themes or even recommendations for further messages. Imagine doing that among not 30, but 300 individuals. That’s 10 winning slogans crowdsourced directly from the groups you need to persuade.

‍

‍

Let’s look again at each focus group stage in a campaign, but this time using Focaldata AI. 

Pre-survey focus groups 🗒️

Using our depth qual platform, the pre-survey focus group is no longer an expensive rubber stamp. 

Because you can launch highly topic-specific in-depth interviews, you can probe the particular areas and terms that you need to clarify before committing them to quant. This does away with the broad-brush, insight-poor, focus groups of typical campaign cycles – which often influence only 2-3 quant questions out of your typical 60 question survey. 

For the cost of a single focus group, you can run 100 in-depth interviews instead. When running campaigns in unfamiliar geographies, these pre-survey focus groups become an invaluable part of understanding your audience and designing a laser-focused survey.

‍‍

Audience deep-dives 🔍

Our qual depth plaftorm transforms the process of bringing segmentations to life. Using our platform targeting capabilities, we find it far easier than traditional qual recruiters to recruit individuals in each segment. This is no surprise: we’ve got far more profiling information on each individual on our platform than a recruiter looking for participants in the wild. With all your target audiences at your fingertips, this means you can interview your most important opinion leader or swing voter in a matter of hours, not weeks. 

We also make it far easier to extract insights quickly. Long gone are the days where you need to wait for the focus group briefing to take the next big decision. Instead, you can simply ask questions of your data to understand what your audiences are thinking. 

Building rich, actionable segment profiles using real-world qualitative data has never been easier.

‍

Message testing 🎯

We’re really excited about the ways AI is completely upending the traditional message testing playbool. In fact, message testing is the use case where the ‘generative’ capabilities of AI reach their full potential. We can use live participant feedback to feed, iterate and improve a message to reach the perfect message for each group. Our research team call this ‘inverted message testing’.  

In traditional campaigns, message testing is usually stop-start. Messages are iterated or twice, tested in a focus group and then the best one is chosen through A/B testing. Over the course of the campaign, you may have tested 10 different iterations of a message, but none the wiser about how the message has performed among different groups or how to improve them. This is where traditional campaign teams get really unstuck by the insight trade-offs presented by traditional quant and qual. 

Using our qual depth AI platform, message testing moves from a start-stop motion to a continuous and cumulative insight loop. You start with a message and use 100s of interviews to refine the message, testing many different versions of the message in the process. The result is an extremely powerful insight loop which can create not one message, but 100s of individually perfect messages tailored to the priorities and preferences of all audiences. 

Qual-level messaging insights, at quant scale. 

‍

Case Study: Focaldata AI powers an 8-week winning political campaign in Chicago

‍

The best part is that we haven’t made these numbers up. They’re taken from a real-world case study where we applied quant platform and qual AI platform in a winning political campaign in Chicago.

Want to see how we did it? You can download the full case study below.