We’re halfway through the week, Reader, how’s your marketing strategy holding up? Today’s tips and insights will help you fine-tune your approach.

In today’s edition:

• Silence kills Reels: Trim your way to retention

• Facebook’s ad algorithm is literal. Are your instructions clear?

• Using Gemini Deep Research for strategic planning

• 🗞️ Industry news from Instagram, OpenAI, Youtube, and more

Trim Your Way to Higher Instagram Reels Views

If you've been wondering why your talk-to-camera Reels aren't getting the views you hoped for, Shannon McKinstrie has a tip: proper trimming.

In this Instagram post, she explains how dead air and awkward pauses act as exit ramps for viewers, killing your watch time metrics. The solution? A simple three-step trimming process that eliminates silence and "ums" to keep viewers engaged.

She walks through exactly how to trim your Reels (starting with dragging until your "mouth is half open" for zero silence!), plus mentions time-saving AI tools that can do this automatically.

A quick read that could dramatically boost your Reels performance - definitely worth checking out if you're serious about growing your Instagram presence! Read more here. Read more here.

Meta Ads Simplified: Less Tinkering, More Converting

Could an overcomplicated Meta ads strategy be why you're getting low-quality leads or serving ads to unexpected audience demographics?

Jon Loomer's latest piece cuts through the confusion with his refreshingly straightforward Common Sense Playbook.

His core philosophy is brilliantly simple: "The algorithm is literal." It does exactly what you tell it to do—showing your ads to people most likely to take your specified action.

He challenges our obsession with targeting, arguing that most inputs are merely "suggestions" to the algorithm now, especially with Advantage+ Audience enabled. That might feel uncomfortable, but it's the future of Meta advertising.

He advocates for a simplified approach—fewer campaigns, consolidated budgets, and focusing on what actually impacts performance: creating compelling ads, irresistible offers, and high-converting landing pages.

The approach isn't about being hands-off but understanding when to trust the algorithm and when to guide it. It's about working with Meta's system rather than fighting against it.

Want to stop wasting money and start getting results? Read more here.

How Are Your 2025 Marketing Goals Going?

If you want to exceed your marketing goals this year, an All-Access Ticket to Social Media Marketing World is your best investment.  

✅ Discover new strategies that improve your marketing results.

✅ Fortify your marketing plans for 2025 and beyond. 

✅ Become much more valuable to your company or clients. 

Get advanced training on marketing strategy, AI, and so much more at Social Media Marketing World 2025–taking place March 30 - April 1. 

Plus, if you secure your All-Access Ticket by this Friday, you'll save $300.

I want to hit my goals in 2025!

AI Assistants: How to Use Google Gemini Deep Research to Develop AI Assistant-Training Documents

Should you be using an AI assistant in your business? If you've ever wished for a consultant who's available 24/7, never gets frustrated, and costs a fraction of a typical advisory fee, then the answer is yes.

Did you know AI assistants can go beyond basic content creation tasks to become strategic partners in your business planning?

How to Use Google Gemini Deep Research for Strategic Planning and Execution

If your business pays for Google Workspace, you already have access to Gemini.

Key advantages of Gemini include:

• Integration with the Google ecosystem

• Very large context window (allowing you to include more information)

• Reference to living Google Docs (which update as you change them)

• Tools like Deep Research for comprehensive information gathering

When you give Deep Research a topic, the system reads your request and comes back with a research plan you can review before proceeding. Allton used Gemini's Deep Research when rebranding his business to focus on helping St. Louis companies.

He asked Deep Research to:

Help me understand the St. Louis business landscape by identifying the top industries in St. Louis, examples of businesses that work in those industries, the challenges those businesses have that AI could resolve, and the associations or organizations that represent those businesses.

The system repeated his requests and added additional research areas it thought would be helpful.

After he approved the research plan, the system scanned over a hundred websites, scraped them and returned a complete report in about 10 minutes. 

This report now informs all of Mike's conversations with his AI assistant, like the brand voice content creation assistant he trained his assistant to keep this focus front and center. 

When he brainstorms an email campaign or a blog post, the assistant constantly reminds him to include St. Louis businesses in the content. This consistent reminder is a strategic safeguard, keeping Mike aligned with his business goals.

Use the RICCE Framework to Create and Train AI Assistants for Various Tasks

While you can use Mike Allton's RICCE framework in individual conversations with AI, creating custom AI assistants allows you to save your instructions for ongoing use. Think of the RICCE framework as training wheels that help you provide all the information an AI needs to serve you effectively.

Role: First, tell the AI what role you want it to assume.

Instructions: Next, clearly explain what you want to accomplish and what you expect from the conversation. Be as specific as possible about your objectives, and instruct the AI to ask you questions if it needs more information. Another important aspect of these instructions is to tell the AI if you want to take a step-by-step approach rather than letting it jump straight to the final product. Without this guidance, the AI will try to be helpful by immediately producing an output.

Today, I am doing an email campaign. I want to reheat my audience because many haven't been opening emails for a while. I want to target them and get them to re-engage with my emails overall. We're gonna go step by step through this process. The first step is to identify who I'm talking to.

Context: Now, you need to provide background information about your business, target audience, and other relevant details that will customize the results to your unique business:

• Who your business is

• Who your target audience is

• Your brand voice and style

• Links to relevant documents or websites

• Strategic planning documents from Deep Research

While you can copy and paste these details into the chat, you can also attach several types of document files.

Google Gemini offers users a helpful enhancement for custom instructions because it can access Google Docs. As you update your brand documents or client information in Google Drive, your AI assistants will always work with the latest version of the data without needing to be retrained.

Constraints: Follow your context with boundaries to help the AI understand your preferences and ensure the output matches your needs. This might include:

• Specific formats you prefer

• Geographic considerations (e.g., "We're only talking to our North American audience")

• Language requirements

• Style constraints (e.g., "Don't overuse emojis")

• Word count limits

For example, if your emails are typically short, you might tell the AI that any email it writes must be 500 words or less. Or you might stipulate long-form content by telling the AI any email it writes must be at least 1500 to 2000 words long.

Evaluation: This final step is to define what success will look like. How will you know if the AI's response is good? Did the AI assistant produce a great email or a ho-hum email? Defining clear criteria for the response helps you evaluate and refine the AI's output.

Today's advice provided with insights from Mike Allton, a featured guest on the AI Explored podcast

Social Media Marketing World Comes to You! 

Attending Social Media Marketing World in person is the best way to experience the conference. But sometimes travel just isn't an option. 

With a Virtual Ticket, you can stream all the keynotes and sessions live. Plus, you'll get access to the recordings of every keynote, session, and workshop.

"My eyes are sore as I have watched way too many sessions today! All awesome stuff. It's like Netflix… just one more," said Julia Doherty.

Isn't it time you experienced the life-changing content from Social Media Marketing World? Grab your Virtual Ticket before Friday to save $400.

I’m ready to attend virtually!

Instagram Blend Test: The platform is testing a new messaging feature that lets you and a friend view a feed of AI-recommended Reels. Source: Mashable

OpenAI Launches Tools to Build AI Agents: The Responses API combines Chat Completions simplicity with tool-use capabilities. Built-in tools include web search for up-to-date information with source citations, file search for retrieving information from documents, and computer use that can automate tasks on computers. The Agents SDK simplifies orchestrating single and multi-agent workflows. Source: OpenAI

Pinterest Labels AI-Generated Content: Pinterest made this change after Futurism published a report highlighting the increasing amount of AI-generated content that is overwhelming some users. Pinterest responded by developing AI detection and labeling features to provide context about what users are seeing. The platform also added a new rule stating they may display a label on image Pins when they detect AI generation or modification, in accordance with the IPTC standard for photo metadata. These tags have been visible in the app for the last couple of months on a small collection of Pins. For advertisements, Pinterest only displays AI content disclosure in the "Why am I seeing this ad?" section rather than directly on the image. Source: Social Media Today

YouTube Testing Ad Suitability Reviews Improvements: The platform is experimenting with automatically sending videos marked "Limited or no ads" for additional review. This may include reviewing newly uploaded private videos. Some monetization decisions could take up to 24 hours. The test is rolling out to a small percentage of creators but will eventually expand to all creators who monetize with ads. Source: YouTube

YouTube Updates Dream Screen: The upgrade integrates Google DeepMind's newest video generation model, V2, making it more powerful. Users can now generate stand-alone 6-second video clips or stitch together multiple clips. The feature also allows the creation of video backgrounds to complement other assets and enhance creative vision. These features are currently available in the US, Canada, Australia, and New Zealand. Source: YouTube

Did You Know?

The term "penny" is not the official name of the United States’ lowest value coin; it's actually called a "cent" or "one cent coin".