AI Marketing: Hype or ROI? A Florist’s $30K Test

The rise of AI answers is reshaping how we approach marketing, promising more personalized and efficient campaigns. But is the hype justified, or are we just chasing another shiny object? Let’s dissect a recent campaign we ran to see if artificial intelligence can truly deliver better results.

We recently wrapped up a three-month campaign for “Bloom Local,” a fictional Atlanta-based florist specializing in sustainable, locally-sourced bouquets. Bloom Local was struggling to stand out against the big national players like 1-800-Flowers and needed a cost-effective way to boost online orders and brand awareness within the metro area.

Our strategy centered around leveraging AI to hyper-personalize the customer journey, from initial ad exposure to post-purchase follow-up. The goal? To prove that AI wasn’t just a buzzword but a genuine tool for driving ROI. We focused on two key areas: dynamic creative optimization and personalized email marketing.

Campaign Overview

The campaign, which ran from March to May 2026, had a total budget of $30,000. We allocated $20,000 to paid social (Meta and Nextdoor) and $10,000 to email marketing and AI-powered personalization tools.

Target Audience: Atlanta residents (ages 25-55) interested in flowers, gardening, home decor, and supporting local businesses. We further segmented this audience based on life events (birthdays, anniversaries), interests (specific flower types, floral design), and location (targeting affluent neighborhoods like Buckhead, Midtown, and Virginia-Highland). We even included those near major hospitals like Emory University Hospital and Piedmont Hospital.

Platforms:

• Meta Ads: We ran targeted ads on Meta, focusing on image and video ads showcasing Bloom Local’s unique floral arrangements.
• Nextdoor Ads: Given Bloom Local’s focus on local sourcing, Nextdoor proved to be a valuable channel for reaching hyper-local audiences within specific Atlanta neighborhoods.
• Email Marketing: We used an AI-powered email platform (think Klaviyo on steroids) to create personalized email sequences based on user behavior and preferences.

Dynamic Creative Optimization (DCO)

Our first step was to implement dynamic creative optimization (DCO) on Meta. DCO allows you to upload multiple versions of your ad creative (images, headlines, descriptions, call-to-action buttons) and let the platform’s AI algorithms automatically assemble the best-performing combinations for each individual user. It’s not a set-it-and-forget-it strategy, though. You still need great raw materials.

We created five different ad variations, each highlighting a different aspect of Bloom Local’s brand:

1. Sustainability: Focused on Bloom Local’s commitment to eco-friendly practices and locally-sourced flowers.
2. Artistry: Showcased the unique designs and artistic flair of Bloom Local’s florists.
3. Convenience: Highlighted the ease of ordering online and fast delivery options.
4. Special Occasions: Featured bouquets designed for birthdays, anniversaries, and other celebrations.
5. Local Connection: Emphasized Bloom Local’s ties to the Atlanta community and support for local farmers.

We also tested different headlines, descriptions, and call-to-action buttons, such as “Shop Now,” “Order Today,” and “Send Flowers.”

Results:

After the first month, the DCO ads significantly outperformed our previous static ad campaigns. The click-through rate (CTR) increased by 15%, and the conversion rate (orders placed) improved by 10%. The AI was clearly learning which combinations resonated best with different audience segments. For example, users interested in gardening responded well to ads emphasizing sustainability, while those celebrating special occasions were more likely to click on ads featuring celebratory bouquets.

Stat Card: Dynamic Creative Optimization (Meta)

Impressions: 1,200,000
CTR: 1.2% (vs. 1.05% for static ads)
Conversions: 480
Cost per Conversion: $38
ROAS: 3.2x

Personalized Email Marketing

Next, we focused on personalizing the email experience. We integrated Bloom Local’s website and CRM with an AI-powered email platform that could track user behavior, such as website visits, product views, and past purchases. This allowed us to create highly targeted email sequences that delivered the right message to the right person at the right time.

Here’s how it worked:

1. Welcome Series: New subscribers received a personalized welcome email introducing Bloom Local’s brand story and highlighting their commitment to sustainability.
2. Abandoned Cart Recovery: Users who added items to their cart but didn’t complete the purchase received a series of emails reminding them of their items and offering a special discount.
3. Post-Purchase Follow-Up: Customers who placed an order received a thank-you email with care tips for their flowers and a request for a review.
4. Personalized Recommendations: Based on past purchases and browsing history, users received personalized recommendations for other floral arrangements they might like.

The AI also helped us optimize the email subject lines and send times to maximize open and click-through rates. It identified patterns in user behavior and adjusted the email schedule accordingly. For instance, we discovered that customers in the Morningside neighborhood were most likely to open emails in the early evening, while those in Downtown Atlanta preferred receiving emails during their lunch break.

Results:

The personalized email sequences delivered impressive results. The email open rate increased by 20%, the click-through rate improved by 15%, and the conversion rate (repeat purchases) increased by 8%. We saw a significant lift in customer engagement and loyalty.

Stat Card: Personalized Email Marketing

Emails Sent: 80,000
Open Rate: 35%
CTR: 5%
Conversions: 280
Cost per Conversion: $35
ROAS: 3.8x

What Worked and What Didn’t

What Worked:

• DCO on Meta: The dynamic creative optimization significantly improved ad performance by delivering personalized messages to different audience segments.
• Personalized Email Sequences: The AI-powered email platform enabled us to create highly targeted email sequences that drove engagement and conversions.
• Hyper-Local Targeting on Nextdoor: Reaching specific neighborhoods with relevant messaging proved effective in driving local orders.

What Didn’t:

• Initial Audience Segmentation: Our initial audience segmentation was too broad, resulting in some wasted ad spend. We refined our targeting based on early campaign data.
• A/B Testing of Email Subject Lines: While the AI optimized send times, A/B testing different subject lines proved less impactful than expected. We focused more on personalizing the email content itself.

Optimization Steps

Based on the initial campaign data, we made several optimization steps:

• Refined Audience Targeting: We narrowed our audience targeting on Meta based on demographics, interests, and behavior. We also excluded users who had already made a purchase from Bloom Local to avoid wasting ad spend.
• Improved Ad Creative: We replaced underperforming ad variations with new creative based on the insights we gained from the DCO testing. We also experimented with different ad formats, such as carousel ads and collection ads.
• Enhanced Email Personalization: We added more personalized elements to our email sequences, such as using the customer’s name in the subject line and including images of flowers they had previously purchased.

We also integrated a chatbot on Bloom Local’s website to provide instant customer support and answer frequently asked questions. The chatbot was trained on Bloom Local’s product catalog and customer service FAQs, enabling it to handle a wide range of inquiries. This freed up Bloom Local’s staff to focus on more complex customer issues.

With the increasing importance of FAQ optimization for marketing, we also updated Bloom Local’s website to address common customer questions.

Final Results

After three months, the campaign exceeded our initial goals. Bloom Local saw a significant increase in online orders, brand awareness, and customer engagement. The overall ROAS for the campaign was 3.5x, exceeding our initial target of 2.8x. The AI-powered personalization strategies played a crucial role in driving these results.

Campaign Summary:

Budget: $30,000
Duration: 3 Months
Total Conversions: 760
Cost per Conversion: $39.47
Overall ROAS: 3.5x

The Human Element Remains Key

It’s tempting to think AI is a magic bullet, but let’s be real: it’s a tool. A powerful one, but still just a tool. I had a client last year, a law firm near the Fulton County Courthouse, that went all-in on AI-driven content creation. They figured they could churn out blog posts about O.C.G.A. Section 34-9-1 (workers’ compensation) and rank overnight. It was a disaster. The content was technically accurate but completely devoid of personality and nuance. Nobody wants to read legal advice from a robot, especially when they’re dealing with something as personal as an injury claim.

Here’s what nobody tells you: AI can analyze data and identify patterns, but it can’t replace human creativity and empathy. It’s up to us, as marketers, to use these insights to craft compelling stories and build meaningful relationships with our customers. AI can suggest the best time to send an email, but it can’t write the email itself (at least, not yet).

Understanding semantic SEO and intent-based marketing is crucial for ensuring AI-driven content resonates with your target audience.

The Future of AI in Marketing

The Bloom Local campaign proves that AI can be a valuable asset for marketers, but it’s not a replacement for human expertise. As AI technology continues to evolve, we can expect to see even more sophisticated applications in areas such as predictive analytics, customer segmentation, and content creation. The key is to embrace these tools while staying true to the core principles of marketing: understanding your audience, crafting compelling messages, and building genuine relationships.

The most exciting part? AI is getting smarter. According to eMarketer, AI-driven ad spending will continue its rapid growth trajectory, and I believe that number will continue to increase as the technology becomes even more accessible.

Will AI answers completely transform the marketing industry? Maybe not entirely. But it’s definitely changing the game, and those who adapt and learn to harness its power will be the ones who thrive. So, what are you waiting for? Start experimenting with AI today and see how it can help you achieve your marketing goals.

To further enhance your marketing strategy, consider focusing on answer targeting to improve conversions.

My advice? Don’t wait for the perfect AI solution. Start experimenting now. Even small steps can yield big results. The Bloom Local campaign showed us that even a modest investment in AI can significantly improve marketing performance. The key is to start, learn, and adapt.