Most cyclists spend five minutes on their power data after a ride. They check average power, feel vaguely satisfied or disappointed, close the app, and learn nothing.
That's a missed opportunity. Your power file from a single ride contains more useful information than most people realize. Once you know what to examine, each ride becomes a source of feedback that actually changes how you train.
Start with these four numbers
Before you dig into graphs, four numbers tell you 80% of what you need to know about any ride.
Average power. The raw average of every second of the ride, including coasting. Useful for understanding total energy output but limited as a training quality metric. A 2-hour ride with lots of descending and stoplight coasting can show an average power of 160W when you spent most of the real effort above 220W. Don't rely on this number alone.
Normalized Power (NP). A weighted average that accounts for the non-linear cost of variable efforts. It's designed to represent the metabolic stress your body actually experienced, not just the raw average. A ride with surges, climbs, and recovery sections will have an NP meaningfully higher than its average power. NP is the more accurate number for comparing training load across different types of rides.
Intensity Factor (IF). NP divided by your FTP. An IF of 0.75 means the ride was at 75% of your threshold. IF gives you a single number to categorize ride intensity. Easy Zone 2 rides land at 0.60-0.70. Tempo rides hit 0.75-0.85. Threshold work sits at 0.85-0.95. Anything above 1.0 is above-threshold and can't be sustained for long.
Training Stress Score (TSS). A combined measure of ride intensity and duration. TSS is how you add up total weekly training load and manage stress week to week without guessing. A 2-hour ride at 0.75 IF produces a TSS of 113. Track this number across weeks and the picture of your training becomes much clearer.
Normalized power: the number that actually matters
Here's why NP matters more than average power for most training decisions.
Your body doesn't experience a one-hour ride at a steady 200W the same way it experiences an hour averaging 200W but ranging from 120W to 320W the whole time. The variable ride is metabolically more demanding, even if the numbers look identical on paper. That's because sustained hard efforts cost more than the linear math would suggest.
NP captures that. It uses a formula that weights higher powers more heavily. A ride with lots of surges and punchy climbs will have an NP 10-20W higher than its average power. A super steady flat TT will have an NP almost identical to its average.
Compare the two on your own rides and you'll quickly learn how variable your efforts actually are. Criterium and group ride power files look very different from long solo training rides. Both can produce the same TSS. The physiological demands aren't the same.
Intensity Factor and your training distribution
Your IF tells you where a ride sits in the spectrum. And patterns in your IF over weeks tell you whether you're training well or just riding around.
Rides below IF 0.70 are genuinely easy Zone 2. If you're chasing polarized training and most of your rides aren't landing here, your easy days are too hard. That's the most common training mistake in amateur cycling at every level, and it's invisible unless you're checking.
IF 0.70-0.80 is the gray zone. Metabolically expensive without providing the specific stimulus of threshold or VO2max work. You're getting the fatigue cost of hard training with the adaptation rate of easy training. This is where wasted fitness lives.
IF 0.80-0.95 is threshold territory. Your FTP-building zone. Productive but high-cost. Requires real recovery afterward.
Log your IF for every ride for a month and look at the distribution. If 40% of your rides are clustering in the 0.72-0.82 range, you're spending too much time in the moderate zone. This is exactly the pattern that feels like hard work while producing slow results.
Reading the power curve
The power duration curve shows your best power across every duration, from a 1-second sprint to your longest sustained effort.
It's the most information-dense thing in your data. A few things to look for:
Where is your curve highest relative to your rider type? Climbers and time trialists typically have relatively flat curves: strong FTP, moderate sprint. Sprinters and crit racers show a steep drop-off from big short-power numbers to lower sustained values. Neither is better. They're different strengths built for different events.
Where is your curve weakest? Most cyclists underperform in one specific zone relative to the rest of their profile. Some are strong at 5-minute efforts (VO2max range) but weak at 20-minute efforts (threshold). Others have a big sprint but 1-minute power falls off quickly. Your weakest zone is usually your best training target.
Is the curve moving up over time? Load the previous 6 months and compare. If your 20-minute power is going up, your FTP is improving. If your 1-minute power is dropping while your 20-minute holds, your sprint is detrained but your aerobic base is solid. These movements show you what's working.
The story in the power graph
Pull up the power graph for any of your rides and look at the shape.
A well-paced solo effort looks relatively smooth. Power varies, but not wildly. You don't see a massive spike in the first 10 minutes followed by a slow erosion. The effort is controlled.
Most people's ride data tells a different story. Look at your threshold intervals specifically. If you're doing a 3x12-minute threshold workout and the power graph shows you starting each interval 20-30W high then fading to finish below target, you're pacing them wrong. Start conservative. Even pacing or a slight negative split is almost always better than going out hard.
And if your average power across the three intervals drops significantly from first to last (say, 258W, 244W, 226W), fatigue is compounding. Either the efforts are too hard, the recovery is too short, or you aren't arriving at the session fresh enough. All three show up clearly in the data once you know what you're looking for.
Time in zones: the weekly picture
Single-ride power data is useful. Looking at your time in zones across a full week is more useful.
Add up how much time you spent in each power zone last week. Most cycling apps show this automatically.
The pattern you want (assuming polarized training) is heavily weighted toward Zone 1-2, minimal time in Zone 3, and a meaningful chunk in Zones 4-5. Maybe 80% easy, 5% moderate, 15% hard. That's the general target.
What most cyclists actually see is a lot of Zone 3. Not enough Zone 1-2. Not quite enough Zone 4-5. That gray-zone distribution feels like hard work. The data says it's producing slow results.
If your distribution is off, the fix isn't complicated. Make your easy days actually easy. And when you go hard, go hard enough for Zones 4-5 to actually show up in your weekly totals.
What to look for after a race
Post-race data tells you things you can't feel in the moment.
Check your 20-minute best from the event and compare it to your tested FTP. If it's higher, your FTP is outdated. Good problem to have. Update it and recalibrate your zones.
Look at how your power faded across the race. If your last 20 minutes is significantly lower than your first 20 minutes, one of three things happened: you started too hard, you didn't eat enough, or your fitness for that duration is lagging. All three are fixable with targeted training.
Compare your IF for the race versus training rides of similar duration. If you're consistently producing an IF of 0.88 in training but only managing 0.82 in races, something is costing you on race day. Pacing anxiety, positioning, nervous energy expenditure before the gun. Worth figuring out.
The number to actually track week to week
If you only follow one metric, make it your weekly TSS trend.
A gradual upward trend over 8-10 weeks means you're building load systematically. A sudden spike of more than 10-15% week over week is a recipe for accumulated fatigue or injury. A sudden drop means you're undertraining or backing off without realizing it.
This is what training periodization is built on at its core: building load deliberately, cutting it back at the right time, then building again. Your TSS trend across weeks and months shows whether you're actually doing that or just riding by feel.
Most cyclists are riding by feel. The data is right there. It just needs someone to look at it.
If you're using Nivvy, the Training Score synthesizes a lot of this into a single weekly grade. But the individual metrics behind that score tell you specifically what to fix. A low intensity distribution sub-score means too much time in Zone 3. A low consistency sub-score means skipped sessions. A low load management sub-score means you spiked TSS too aggressively. The score is the summary. The underlying power data tells you what to actually change.