When push comes to shove, anecdotal evidence is more persuasive than statistics. That applies most to decisions that affect you directly — the types of decisions you make every day at the C-suite and VP levels.
Incomplete information, biases, and rumors can be persuasive. There’s a place for that — somewhere outside of the decision-making process. Take a few minutes here to learn about a better way to optimize your streaming operations.
- You probably use more anecdotal evidence than you realize.
- Implementing an optimization system can free up your time and get better results.
- Robust, multi-viewpoint benchmarking is the foundation of actionable roadmaps.
- Potential for improvement, audience reach, and engagement impact are the three key factors to look for when optimizing.
Examples of Anecdotal Decisions
Do you make anecdotal decisions? Probably not — at least not entirely.
Leaders quickly learn to hone in on the most important, most reliable information. Still, if you reflect on your optimization workflows, you might be surprised at how much some of the following types of storytelling influence you at one point or another in the process.
Most technical leaders have at least some engineering background. Whether it’s from work, academic background, or your experience using your organization’s service last night, remember that your common sense might not apply. In other words, your experience might not even be relevant to the decision — much less a universal truth.
Word of Mouth and Rumors
Another trap that many tech leaders fall into is putting too much faith in unsubstantiated information from colleagues, pundits, or prominent figures. You can use this knowledge as a guidepost, but it is unreliable as a decision-making tool. Remember to verify and contextualize the statements with your own research and your organization’s data.
Advertising and marketing materials are ubiquitous. Advertisers don’t want to provide objective or even unbiased statements — they’re telling a story that furthers their own business goals. Use this type of information only to the extent that your objectives align with the advertiser’s.
Biased or Incomplete Selections
Biases and incomplete information contribute to anecdotal decisions. Incomplete or cherry-picked data leads to a skewed picture of performance. Confirmation biases can lead to false conclusions or slow reactions, making you ignore the issues right in front of you. Having robust data collection and targeting specific metrics can narrow things down enough for you to focus on what really matters.
Stories motivate people, and that holds true with anecdotal evidence. Just make sure to use relevant data to confirm that the anecdotal aspects of your decisions, if there are any, remain in line with your true objectives. At the end of the day, remember that any situation is testable.
Optimizing Your Optimization Process
The scariest part of anecdotal decision-making is that your choices may work out. The thing to remember is that they won’t work out consistently. For consistency, you need a system.
However, it’s one thing to say that you need a system and it’s another thing to develop one. Here’s an example of a decision-making cycle that has led to consistent, predictable, and manageable optimization for enterprises in the VoD and broadcasting arenas:
- Establish clear business objectives: It almost goes without saying that you need goals. However, you also need clarity and specificity in those goals. For example, if you want to increase revenue, are you aiming for ad revenue or subscriptions?
- Select a set of relevant audience engagement KPIs: Your key performance indicators will be different for each objective. Continuing the example: You would want more plays and more viewer minutes if you wanted to increase advertising revenue.
- Connect QoE metrics to KPIs: This is the main step that makes your optimization process predictable and actionable. You have to connect any business growth or cost reduction to measurable QoE metrics such as video start times, playback failures, bitrates, or buffering times. You do this by linking your relevant QoE metrics to KPIs, which, in turn, you have already linked to your business objectives.
- Benchmark performance: The next step is to determine what’s normal for the QoE metrics you selected. Benchmarking leads progress by balancing realistic expectations and ambitious goal setting. Check out the following section on benchmarks for an in-depth discussion.
- Make a roadmap: Once you know where you stand, it’s time to work out where you’re going. Plot out the path to improving KPIs based on benchmarks, audience data, and current QoE analytics.
- Analyze root causes: Get into the where, what, who, and how of solving any issues.
- Optimize services: This is the execution part of the process. Mobilize your team to resolve the root of your issue, informed by any guidance and insight you developed in the previous steps.
- Validate actions: It’s not over until it’s over. Confirm that you met the technical requirements necessary to achieve your business objectives. Do this through systematic validation of operations.
- Repeat: Set new, clear objectives in line with your organization’s overall vision of success.
Drilling down into key metrics is the only way to make this process work. Focus on the metrics that matter for your goals and your operations, improve performance, confirm the effect, and repeat.
Benchmarking as a Core Optimization Concern
How do you know what’s normal? How do you know what’s possible?
The answer is simple: You measure and compare. Benchmarking is essential if you want to direct your organization’s resources toward practical, impactful goals.
The first step to effective benchmarking is to standardize your measurements. You need apples-to-apples comparisons. Especially if you’re comparing metrics from different sources, you need to execute the same sanitization and standardization processes on all data.
Moving from Standardization to Action
Once you have a standardized point of view, you can start looking at things in a more objective way. You will also probably have a leg up on your competition.
For example, you can see which of your many QoE metrics that impact a KPI are underperforming the industry. From there, you can select the metrics that:
- Have the greatest potential for improvement
- Affect the largest number of audience members
- Impact engagement the most
These are the foundations of an efficient roadmap. Of course, the entire optimization process builds on granular data about the user experience.