The big idea: Data tells you which visuals actually work. Test on purpose, read the numbers honestly, and let performance shape your next creative decision.
See it in action
Creative Scorecard Lab
Three versions of one ad for Trailhead Outfitters. Pick a goal, pick the metric that matters for it, then diagnose what to change visually.
| Metric | A | B | C |
|---|
Key concepts
- The metrics that matter: Reach, engagement rate, saves, shares and click-through rate tell different stories about the same visual.
- Saves and shares: These signal content people found valuable enough to keep or pass on.
- A/B test visuals: Change one element, like the image, color or headline, and compare results.
- Look for patterns: Review your top and bottom performers together and ask what they have in common.
- Data informs, people decide: Numbers show what happened. Your judgment explains why and what to try next.
Frameworks from class
Look past vanity metrics in video
Views alone do not tell you whether a video worked. A video with fewer views but strong engagement often does more for the brand than a viral one people abandon. Track the metrics that show real attention and action.
- Average watch time: how long people actually stay
- Completion rate: the share who watch to the end
- Click-through rate: how many take the action you asked for
- Organic shares: how far it travels past your followers
Tools that show what people do
Some AI-powered tools go beyond platform dashboards. Each answers a different question about how people respond to your visuals.
- Heatmaps and session recordings, such as Hotjar: where people look, click and scroll on a page
- A/B testing platforms, such as Optimizely: which version of a page or visual wins
- Social listening, such as Brandwatch: what people say about the brand and how they feel
Try it
Look at the last ten posts from a brand you follow or manage. Rank them by engagement, then compare the top three and bottom three. What visual patterns do you see? Write one test you’d run next.
Tips and best practices
- Pick the goal before you open the dashboard. Decide whether a post is meant to stop the scroll, earn saves or drive sales, then judge it on the one metric that matches. A high-reach post can still be a failure if the goal was sales.
- Test one visual change at a time. Swap only the image, the headline or the color, and keep everything else the same. If you change three things and results improve, you will not know which one worked.
- Read the first second separately from the rest. Hook rate tells you whether the opening frame stopped people. Hold rate and clicks tell you whether the rest delivered. Fix the part that is actually failing.
- Treat saves and shares as a usefulness signal. People save what they want to come back to and share what they want others to see. Checklists, how-tos and strong opinions tend to earn both.
- Line up your best and worst posts side by side. Look at your top three and bottom three as images, not numbers. Note what the winners share, such as faces, contrast or a clear subject, and turn that into your next test.
- Write down what you learned, not just what won. Keep a simple test log with the change, the result and your best guess at why. Over a semester or a quarter, that log becomes a playbook your team can use.
Quick check
Think through your answer, then select a question to reveal one.
Why do saves and shares matter?
They show people found the content valuable enough to keep or pass along.
What’s the rule for a clean A/B test?
Change only one element at a time.