How Lafayette American turns Mazda’s comment sections into creative briefs
About Lafayette American
Industry
Advertising and Marketing
Founded
2018
Headquarters
Detroit, MI
Before Siftsy
Manual comment review alongside social management and competitive analytics tools
3 hrs → 20 min
To analyze the comments on high-volume TikTok content
10 min
Saved per report by uploading a whole account instead of copying links
How the social team behind Mazda’s TikTok turned comment sections into monthly benchmarks, creative briefs, and a clearer read on a controversial launch.
Challenge
Mazda wanted more than a few top comments
Caite Crocker, Associate Director of Social Strategy at Lafayette American, leads TikTok content strategy for one of the agency's clients, Mazda. She also leads Mazda's community management across social platforms. Emily Carion, a digital strategist, works with her on both and has a big hand in reporting. Of the clients Caite and Emily work with, Mazda has been the most interested in social analytics, and it wants to know what the numbers mean for the content it makes next.
Before Siftsy, the team already had tools for publishing, community management, reporting, and competitor data. They covered performance well, but they didn’t show how people felt about a post.
To understand the reaction to a post, Emily would read the comments herself. She would scroll through, get a sense of the mood, and pull out anything Mazda should know about.
With this method, she'd catch individual comments worth flagging, but it was hard for her to identify what the overall sentiment was or which concerns kept coming up across posts. On a TikTok post with thousands of comments, Emily estimates a proper read would take two to three hours. Doing that for every post in a month wasn’t realistic, so reports usually leaned on a few top comments and the team’s own impression.
Moreover, some of the most important conversations weren’t on Mazda’s account at all. Certain creators post about Mazda’s cars all the time, often without tagging the brand, and the team’s other tools didn’t make those videos easy to evaluate.
Quick questions from the client had the same problem. If a post took off and Mazda asked the next day whether it was still gaining traction, the team could only scan the newest comments and give their best guess.
Solution
Sentiment for every post, not just a few flagged comments
At Lafayette American, comment analysis is a two-person process. Emily brings Mazda’s posts into Siftsy, either a single link or a whole social account at once, and filters out Mazda’s own replies so the results show what viewers think. She groups the posts into collections that match her reports. Her findings then go into the reports that the team sends to Mazda, and into the monthly briefs that Caite writes for the creators who make Mazda’s TikToks.
What Siftsy added, in Emily’s view, was the overall sentiment. Before, she could point to individual comments. Now she can say how a whole comment section feels and compare that across posts, months, and platforms. She uses it for TikTok reporting and for community management reports on every platform the team covers.
What the monthly reports show
Each month, Emily ranks Mazda’s owned TikTok posts by their Siftsy Score and how positive each comment section was. As mentioned earlier, sentiment is usually the headline of the report. The team compares each post with the rest of the month and with earlier months, and those ranges have become a benchmark they plan against.
Over several months, the reports have started to show which car models people respond to most. Emily described it as learning “what models drive more positive conversations on our owned content,” and which ones draw more negative sentiment. For example, Mazda MX-5 Miata posts have consistently come out on top, so when the team wants a warm response, they have a reason to feature one.
When the comments asked for a specific car
In September of 2026, Emily’s report showed a lot of enthusiasm for the Mazda CX-50 in Cypress Green. Caite took that insight straight into the next brief for Mazda’s creators.
Interestingly, the color drew different requests depending on the social platform. Under an Instagram post, people wanted Cypress Green on the Miata. Under a TikTok about the same color, they wanted it on SUVs. For Emily, that was a useful look at which models each audience wanted to see.
Not every idea from a report becomes a video, but Caite now has the audience’s own requests to work from whenever she writes a brief.
Seeing the Mazda CX-5 reaction beyond Mazda’s own posts
Sometimes Emily’s findings go to Mazda instead of the creators. When the 2026 Mazda CX-5 launched, it replaced physical buttons with a large touchscreen, which drew mixed reactions online. Mazda wanted a clear picture of how people were responding.
Mazda’s own posts only showed part of that conversation. People don’t always talk the same way under a brand’s post as they do under someone else’s video. So the team added viral Mazda CX-5 videos to Siftsy, whether or not the creators of those videos worked with Mazda.
Emily could then separate the negative reactions from the positive ones and identify any common threads. She could also compare what people said on Mazda’s own posts with what they said on creator videos, and check whether the same themes were showing up.
From there, she checked whether those themes showed up on other platforms and sent an overview to her Mazda client contact. On Mazda’s side, there was a push to pass some of that feedback along to research and development.
Adding every video from a fast-moving creator at once, and resyncing posts that are still gaining traction
The team also turns to Siftsy when a single creator or post suddenly draws a lot of attention. During an escalation last year, one creator posted a long run of videos involving Mazda. The team’s other tools couldn’t easily evaluate creator content the brand wasn’t tagged in, so they added all of the videos to Siftsy. That gave them reach, engagement, and comment sentiment across the whole set without having to open each video.
A single post can move just as fast. When one takes off, Mazda usually asks the next morning whether it’s still gaining traction. The team then uses the Resync feature in Siftsy to provide an answer with updated numbers and sentiment, instead of guessing from just reading the latest comments.
Results
Recommendations Mazda can trace back to what people actually said
When Emily first added Siftsy findings to Mazda’s reports, the response from the client was so positive that they became a regular part of every monthly and quarterly TikTok report. Emily says the reporting has improved “tenfold.” And when asked to finish the sentence “With Siftsy, we can now…,” she said simply, “understand our community.”
For Caite, the difference is in the recommendations she makes to Mazda and its creator partners. They used to rest on a few standout comments, but now they’re backed by sentiment across the whole conversation.
Caite is already thinking about what comes next. Mazda’s beginning to plan for the next fiscal year, and she expects the sentiment data the team has collected to shape the bigger campaign ideas they bring to Mazda next year.
She has also shown the reports to Lafayette American’s leadership, and the agency is now looking at having more of its strategists use Siftsy across client projects.
