Blog

From Feature Usage to Business Impact: Flipping the Analytics Lens

Écrit par
Nicolas Narbais
5/11/25

Créez des vidéos IA avec plus de 230 avatars dans plus de 140 langues

Créer une vidéo IA gratuite
Commencer GRATUITEMENT
Démarrer

Most product analytics start with features. We ask: who used what, how often, and how does that correlate with churn or growth? But this approach often stops short. It treats activity as a proxy for value without proving that activity actually creates business impact.

What if we flipped the perspective? Instead of starting with usage and trying to infer impact, we could start with the business value our users want and trace it backward into measurable signals.

YouTube approaches platform value by focusing on improving views, impressions, and click‑through rates. Once those goals are clear, the question of how to achieve them becomes the responsibility of the video creation tools that enable content performance.

Step 1. Start From the User’s Business Outcome

Every user buys a product to improve something measurable in their business.

For Synthesia, that might be:

  • Recruiting teams: higher candidate engagement and faster time-to-hire
  • L&D teams: higher training completion and knowledge retention
  • Sales teams: improved outreach response rates
  • Support teams: fewer tickets and faster resolution

These outcomes are the real north star. They are not “features used” or “videos created”, they’re business changes. Everything else should map back to these.

However, those outcomes are not often shared or known by the end user, the next best proxy to assess whether video had a positive impact is consistent usage. But, we can also go further.

Step 2. Define Business Impact Proxies

Most of these outcomes aren’t directly measurable or difficult to correlate with a single video. So the next step is identifying in-product proxies that correlate with business success.

Examples:

  • Number of videos published and actually used (embedded, shared, watched)
  • Percentage of target audience reached (impressions)
  • Average watch time and completion rate
  • Engagement actions such as clicks, reactions, or quiz completions
  • Repeat viewership or returning viewers

If these metrics trend upward, it’s a good signal that the user is realizing business value, which usually predicts retention and expansion.

As an example, would you return to see a video or check the next one if this would not be valuable for you?

Building a compounding system

One key element here is to create a compounding system. Like in social media objectives, we care about growing the audience, views, and impressions from one week to another. With the same effort, how can we have a bigger impact?

In the case of Synthesia, it is difficult to measure complete understanding, and sometimes it may not matter. For example, I can't always verify if someone knows how to use advanced SQL queries. Yes, quizzes exist but they test up to a point.

Another take on it. If people return and find my content valuable, that’s a positive sign. Over time, as more content becomes available, the overall coverage of questions and answers will improve. This is about getting better with the same resources. And for my content on SQL, users will come back as needed.

Step 3. Evaluate Video Performance Metrics

  1. Impressions → Click-through rate (CTR): How discoverable the video is
    1. Is the video displayed in multiple places?
    2. Can people browse through videos?
    3. Is the video surfaced in search queries?
  2. CTR → Watch time: How well it captures attention
    1. Is my thumbnail compelling enough?
    2. Do I have a good brand reputation?
  3. Watch time → Completion rate: How well it holds attention
    1. Is my video interesting? Should I use more animations or a better hook?
    2. Is it translated into the viewer's language?
    3. What is my dropout rate after 30 seconds vs my other videos?
  4. Completion → Engagement actions: How well it drives understanding or behavior
    1. Do I keep the content entertaining enough throughout?
    2. Do I use interactive elements or video agents?
  5. Unique viewers / Subscribers: How addictive is my content as a whole
    1. Do I get more subscriber to my content?
    2. Do users come back to the same video or consumer more of them?

Once you’ve defined what success looks like, zoom in one layer. Now you’re analyzing how effectively each video drives that outcome.

At this level, you’re not talking about features anymore. You’re measuring content effectiveness, whether what users create actually works and then you suggest features to improve those metrics. This is where in most cases Synthesia provide a large set of features to improve those metrics.

Step 4. Connect Feature Usage to Outcomes

Only after you’ve mapped impact and engagement should you look at feature adoption. This is where the data analyst’s job becomes prescriptive: which features or behaviors cause better outcomes?

Example questions:

  • Do users who use screen recorder have higher watch completion rates? Can we improve completion rate with better screen recording?
  • Do custom avatars drive more engagement than stock ones? Could we improve engagement with our latest avatar tech?
  • Do viewers of the sales related content have a higher win rate and smaller sales cycle?

Now you can connect product usage to content performance, audience engagement, and business impact. The chain becomes clear.

Step 5. Build the Causal Chain

Think of this as a hierarchy of influence:

Product Usage → Video Quality → Viewer Engagement → Business Impact

Step 6. Change the Conversation With Users

When you speak to users or stakeholders, lead with impact metrics, and not feature adoption.

Example:

We noticed your videos get fewer impressions per publish than the median. Are you promoting them in multiple channels?

or

Your watch time drops after 45 seconds. Try adding animation or using a custom avatar to keep attention.

and not

You are not using the screen recorder.

This reframes success around outcomes instead of usage compliance. It’s a shift from “use more features” to “get more value.”

Closing Thought

Metrics should not just describe product usage; they should explain why users succeed. When analytics starts with business impact and works backward, it transforms from a reporting function into a growth engine. You stop counting clicks and start understanding value creation.

That’s how you move from usage metrics to impact metrics and from retention reporting to retention design.

Démarrer

Créez des vidéos de qualité professionnelle avec des avatars IA et des voix-off dans plus de 140 langues

Transformez en quelques minutes vos textes, PPT, PDF et URLs en vidéo!

En savoir plus
Créer une vidéo IA gratuite
Créer une vidéo IA gratuite
Écouter

Plus de 50 000 équipes nous font confiance

faq

Foire aux questions