You cross the line, you stop your head unit, you roll through the finish area looking for water and someone handing out bananas. Maybe you feel great. Maybe you feel destroyed. Either way, within about twenty minutes you're going to pull out your phone, upload to Strava, glance at a couple of numbers, and get on with your day.
That file you just uploaded contains months of training direction. And you looked at it for about twelve seconds.
I get it. You're tired. You want food. You want to sit down. The last thing you feel like doing after four hours in the saddle is homework. But this is precisely where the gap opens up between cyclists who keep improving year after year and the ones who stay stuck at the same level despite racing regularly.
The race itself is the test. The analysis is where you actually learn.
Why most amateurs skip analysis (and what they miss)
There's a pattern I see constantly in the Not Done Yet community. Someone finishes a sportive, posts their time in the group chat, gets a few well-done messages, and that's the end of it. No debrief. No structured look at what happened. No connection between the event and the training that comes next.
Why? Three reasons, usually.
First, most riders simply don't know what to look for. They see a wall of numbers in TrainingPeaks or Strava and don't know which ones matter. Second, there's a psychological thing — if the ride went badly, nobody wants to sit with that data. And if it went well, they assume there's nothing to learn. Both are wrong. Third, analysis feels like it requires a degree in sports science. It doesn't. It requires a framework.
Joe Friel has been saying for decades that most amateur cyclists waste their race data. His argument is simple: a race is the most expensive training session you'll ever do — weeks of taper, travel, entry fees, emotional energy — and then you extract almost nothing from it afterwards. That's a terrible return on investment.
Tim Kerrison's approach at Team Sky was the opposite extreme. Every stage of every Grand Tour was dissected. Power files, heart rate files, nutritional intake, sleep data, even rider mood. The famous "marginal gains" philosophy wasn't just about hexagonal bolts and heated overshorts. It was fundamentally a culture of relentless post-event analysis. Every race produced a list of things to work on.
You don't need a performance team. You need about 45 minutes and a system. The good news is the system is the same every time, and once you've done it three or four times, it becomes second nature.
The 24-hour debrief — write notes while fresh
Before you touch your power file, you need to capture what the numbers can't tell you. And you need to do it within 24 hours, because the subjective experience of a race degrades surprisingly fast.
Here's what I mean. Three days after a sportive, you'll remember that you struggled on the final climb. But you won't remember that you'd been sitting in the wind for fifteen minutes before that climb because you lost the group on a roundabout. You won't remember that you fumbled your gel wrapper at the feed station and missed an entire intake. You won't remember that your left shoe felt tight from kilometre 60 onwards and you were adjusting your foot position every few minutes. Those details matter, and they vanish.
The debrief doesn't need to be fancy. Open the notes app on your phone and answer these questions:
Before the start: How did you feel warming up? Sleep the night before? Breakfast timing and content? Any anxiety, any niggles, any equipment issues?
First quarter: How was the pacing? Did you get caught up in the fast start? When did you settle into your own rhythm? How was the group you were in?
Middle half: Where did things feel good and where did they start to slide? Any mechanical issues? Did you eat and drink on schedule? Where were the hard patches — climbs, headwinds, surges?
Final quarter: How did you fade (or not)? Where did you lose or gain time relative to the riders around you? What was your mental state in the last 30 minutes? Did you have anything left at the finish?
Overall: If you could do the event again tomorrow with one thing different, what would it be?
That last question is gold. It forces you to prioritise. Write all of this down before you look at any data. The reason is important: if you look at the numbers first, you'll unconsciously rewrite your memory to match them. Dr David Dunne talks about this in the context of evidence-based training — the subjective experience and the objective data need to be captured independently before you compare them, otherwise you get confirmation bias in both directions.
Power file analysis — NP, IF, VI, what they tell you
Right. Now you can open the file. If you're using TrainingPeaks, WKO, Intervals.icu, or even Golden Cheetah, the key metrics are the same. You're looking at four numbers first.
Normalised Power (NP)
This is the single most important number in your race file. Not average power. Normalised Power.
Average power is a liar. If you ride 200 watts for an hour, your average is 200W. But if you alternate between 300W surges and 100W coasting, your average is also 200W — and your body knows full well those two rides were not the same. The surges cost you far more glycogen, created far more fatigue, and did far more muscular damage.
Normalised Power accounts for this by applying a rolling 30-second average and then raising it to the fourth power. The effect is that surges are weighted much more heavily, which is exactly what happens physiologically. Your NP from an event tells you the true metabolic cost of the ride.
For a sportive, compare your NP to your current FTP. If your NP was 85% of FTP for a 4-hour event, that's aggressive. If it was 72%, you probably had more in the tank. If it was 91%, you either had an exceptional day or you're going to see some ugly numbers in the second half of the file.
Intensity Factor (IF)
This is simply NP divided by FTP. It converts your effort into a ratio.
| Event Duration | Sustainable IF Range | |---|---| | 1 hour (crit, TT) | 0.95 – 1.05 | | 2 hours | 0.85 – 0.95 | | 3 hours | 0.78 – 0.87 | | 4 hours | 0.72 – 0.82 | | 5+ hours | 0.65 – 0.78 |
If your IF was above the top of the range for your event duration, one of two things happened: your FTP is set too low (test it), or you overcoked the effort and there will be evidence of that in the pacing data. If it was below the bottom of the range, you left something on the table.
This is incredibly useful information. It tells you, objectively, whether you raced within your physiological capacity.
Variability Index (VI)
This one is underused. VI is NP divided by average power. A perfectly even effort has a VI of 1.00. Every time you surge above your average and then recover below it, the VI climbs.
For a road race, a VI of 1.10 to 1.20 is normal because road racing is inherently variable — surges, attacks, corners. For a sportive, where you mostly control your own pace, a VI above 1.05 is a red flag. It means your pacing was erratic, and erratic pacing always costs more energy than it should.
Here's where it gets really interesting. If your VI is high and your IF was also above range, you have a clear diagnosis: you surged too much and burned through your fuel. If your VI is high but your IF was reasonable, you paced badly but had the fitness to absorb it. Both are fixable — but the fixes are completely different.
Training Stress Score (TSS)
TSS tells you the total load of the event. It's calculated from IF and duration. A TSS of 100 equals one hour at FTP. For most sportives, you'll be somewhere between 180 and 350 TSS depending on duration and intensity.
The number itself isn't as useful for race analysis as it is for recovery planning. A TSS above 300 means you should plan for at least two easy days before any quality training. Above 400 and you're looking at 3-5 days of reduced load. This matters because the biggest post-race mistake I see is riders feeling motivated after a good event and jumping straight back into hard training. Your body doesn't care how motivated you are. It needs time.
| TSS Range | Recovery Required | |---|---| | Under 150 | Normal training next day | | 150 – 250 | Easy day, then resume | | 250 – 350 | 2-3 easy days minimum | | 350 – 450 | 3-5 days reduced load | | 450+ | Full recovery week |
Heart rate and RPE overlay — aerobic decoupling, fade analysis
Power data tells you what you did. Heart rate data tells you what it cost your body to do it. When you overlay the two, patterns emerge that neither shows alone.
Aerobic decoupling
This is one of the most revealing metrics you can pull from a race file, and most riders have never heard of it.
Split your ride into two halves by time. Compare the ratio of average heart rate to average power in each half. If your heart rate rose by more than 5% relative to your power output in the second half, you experienced significant aerobic decoupling. Your cardiovascular system had to work progressively harder to maintain the same output.
A little bit of decoupling (2-3%) is normal over long events due to cardiac drift, dehydration, and heat. But anything above 5% is a strong signal. It means your aerobic base — your ability to sustain sub-threshold work over long durations — was the limiter. Not your FTP. Not your VO2max. Your base.
The practical implication is massive. If decoupling was your issue, the fix is more Zone 2 volume, not more threshold intervals. Most self-coached riders get this backwards. They see themselves fade in a long event and think "I need to be fitter." What they mean is "I need a higher FTP." What they actually need is more time at low intensity building the aerobic engine that supports everything above it.
Joe Friel's test for aerobic base fitness is essentially this: can you hold a steady power-to-heart-rate ratio for 2-3 hours? If not, that's where the work is.
The fade profile
Beyond the numbers, look at your power curve across the event. Create time-based segments — every 30 minutes, or every major section of the route — and compare average power across them.
A common pattern: the first segment is 10-15% above the overall average. The last two segments are 10-20% below. That's a rider who went out too hard and paid for it, and the numbers confirm what the legs already told them.
A less common but important pattern: power holds steady until one specific section, then drops sharply. That's not a pacing issue — that's a limiter. Maybe it was the first big climb. Maybe it was after a missed feed. Maybe it was at the 3-hour mark regardless of terrain. That specificity gives you something to target.
RPE comparison
Now pull out your written debrief. Compare what you wrote about how you felt with what the data shows. Where do they agree? Where do they disagree?
If you felt great but the data shows you were fading, you might be under-calibrated on RPE — you didn't notice the decline because it was gradual. That's common and it's fixable with more attention to your numbers during training.
If you felt terrible but the data shows you held steady, the issue was psychological, not physiological. That's a different kind of work entirely. Tim Kerrison was known for separating perceptual data from performance data precisely for this reason — a rider might feel awful on a stage and still produce world-class numbers, and knowing that distinction matters for what you train next.
Pacing review — did you go out too hard?
This gets its own section because it's the single most common error in amateur racing and sportives. And it's fixable.
Pull up the first 20% of your ride and compare it to the overall. If your power in that opening segment was more than 10% above your overall NP, you went out too hard. Full stop.
The physiology is clear enough. Those early surges above threshold deplete glycogen at a disproportionately high rate because you're relying on anaerobic metabolism. You feel fine at the time because you're fresh. But you've already written a cheque your body will cash later, usually around the 60-70% mark of the event when the glycogen deficit catches up with you.
Dr David Dunne has talked about this in terms of the glycogen cost of supra-threshold work — it's not linear. Ten minutes at 120% of FTP costs far more glycogen than ten minutes at FTP, despite only being "20% harder." The relationship is exponential, and it's why pacing discipline in the first hour of a long event is worth more than any amount of fitness.
The climb-by-climb audit
If your event had significant climbing, break out each major ascent separately. Compare your power on the first significant climb to the last one. Compare your average power on climbs to your average power on the flats.
A well-paced rider will show roughly even power across climbs, with maybe a slight decline on the final one. An overpaced rider will show a dramatic drop-off. I've seen files where the first climb was ridden at 95% of FTP and the last at 68%. Same rider, same legs, same bike. The difference was all in the first hour.
If you find this pattern in your data, the fix is brutally simple: ride the first climb of your next event 10 watts below what feels comfortable. You'll hate it at the time. You'll love it at the finish.
Negative splitting
The gold standard for sportive pacing is a very slight negative split — the second half marginally faster or more powerful than the first. Almost nobody achieves this naturally because it requires the discipline to feel like you're going too easy at the start. But it's what Tim Kerrison demanded of his Grand Tour riders for individual time trials, and it's what the research consistently supports for steady-state efforts.
Check whether you've ever achieved a negative split in any event. If not, pacing is a skill you need to practise in training, not just know about in theory. Set a target of riding your next long training ride (3+ hours) with the second half at 3-5% higher average power than the first. It's harder than it sounds.
Nutrition audit — did your fuelling plan survive contact with the event?
Every rider should go into an event with a fuelling plan. Most do. The question is whether they actually executed it.
The nutrition audit is simple: go through your pre-ride notes (you did write a fuelling plan, didn't you?) and compare what you planned to eat and drink with what you actually consumed.
Carbohydrate per hour
This is the number that matters. Current sports science supports 60-90g of carbohydrate per hour for events lasting over 2.5 hours, with some riders tolerating up to 120g/hr using glucose-fructose combinations. What matters is whether you hit your target consistently through the event.
Count your gels, bars, drink mix, and any food you consumed. Add up the total carbohydrate and divide by the number of hours. Compare that to your plan.
| Planned Intake | Actual Intake | Diagnosis | |---|---|---| | 80g/hr | 80g/hr | Nailed it | | 80g/hr | 60g/hr | Under-fuelled — likely faded late | | 80g/hr | 40g/hr | Major deficit — almost certainly bonked | | 80g/hr | 90g/hr | Over-fuelled — possible GI issues |
The most common pattern I see: riders start the event fuelling on schedule, then stop eating around the 2.5-3 hour mark because they feel full, nauseous, or just forget. By hour 4, the deficit catches up and they fade. The data often shows this clearly — power drops at exactly the point where the fuelling gap would have become physiologically relevant.
Hydration and electrolytes
Harder to audit precisely, but look at your bottle count. How many bottles did you go through? Was that consistent with your plan? If you finished the event with a full bottle on your bike, you under-drank. If you ran out before the last feed station, you didn't carry enough.
For most riders in temperate conditions, 500-750ml per hour is the target range. In hot conditions, 750-1000ml. Dehydration of even 2-3% of body mass impairs power output measurably, and it's one of the most common limiters in summer sportives.
GI distress
This is where the written debrief becomes critical. If you experienced any stomach issues — bloating, nausea, urgency — note exactly when in the event it started and correlate it with your intake. Was it after a new product? After eating too close to a hard effort? After mixing products that don't combine well?
GI problems in events are almost always fixable. They're either a product tolerance issue, a timing issue, or a concentration issue (too many grams dissolved in too little fluid). But you have to diagnose which one before you can fix it, and that requires the combination of your written notes and your intake log.
Tactical review — positioning, group dynamics, when you gained or lost time
This section applies more to road races than to sportives, but even in a sportive, tactics matter.
Group riding efficiency
If your event had significant group riding, look at the difference between your average power and the average speed. Were you drafting efficiently? Did you spend too much time on the front? Did you lose groups at critical moments and have to chase back on alone?
Chasing back to a group after being dropped is one of the most expensive things you can do in an event. A five-minute chase at 110% of FTP can cost as much glycogen as twenty minutes of steady group riding. If your file shows repeated surges above threshold followed by recovery periods, check whether those correspond to getting dropped and chasing back. That's not a fitness problem — it's a positioning problem. Stay in the group, stay in the draft, ride at the front when it's your turn and only when it's your turn.
Where you gained and lost time
Use Strava segments or manual section splits to identify where in the course you were fastest and slowest relative to the field (or relative to your own target time).
Were you losing time on the climbs? The descents? The flat sections? Into headwinds? This is incredibly revealing because it tells you specifically what kind of fitness or skill was the limiter.
If you consistently lose time on climbs relative to the flat, your power-to-weight ratio needs work. If you lose time on the flat, your sustained power at threshold needs work. If you lose time on descents, it's a skills issue — bike handling, cornering confidence, or both. Each of these leads to a completely different training emphasis.
Decision points
Think about the moments in the race where you had a choice. When the pace went up, did you follow or let it go? When a break formed, were you in it? When you hit the feed zone, did you stop or grab and go?
These aren't just tactical questions. They reveal your psychology under pressure. A rider who always lets the group go at the first surge is making a habit of that decision. A rider who always chases the break and then fades is making a different kind of mistake. Both patterns are fixable, but only if you notice them.
Building an action list — turning data into training priorities
This is the point of the entire exercise. Everything above is diagnosis. This section is the prescription.
After you've done the full analysis — written debrief, power file metrics, heart rate overlay, pacing review, nutrition audit, tactical assessment — you should be able to identify three to five specific things that limited your performance.
Not vague things. Not "get fitter." Specific things.
Here's what a good action list looks like:
-
Aerobic decoupling was 8% — prioritise Zone 2 volume. Add one extra endurance ride per week for the next training block, targeting 2.5+ hours at 65-72% of FTP. Track Efficiency Factor over the block to measure progress.
-
Under-fuelled by 20g/hr in the final two hours. Set up a timer on the head unit for every 20 minutes. Pre-open gel wrappers. Practise the fuelling plan in every long training ride for the next four weeks.
-
First climb was 12% above event NP — overpaced the opening. In the next three long rides, deliberately ride the first 45 minutes at 5% below target power. Build the discipline of starting easy.
-
Lost two minutes on descents relative to comparable riders. Book a skills session or dedicate one ride per week to technical descending practice on familiar roads.
-
VI was 1.08 in a sportive — too much surging. Train with power smoothing awareness. Practise climbing at steady-state power rather than matching the riders around me.
See the pattern? Each action item has a diagnosis (what went wrong), a specific intervention (what to do about it), and a measurable way to track whether it's working.
Here's where most riders go wrong: they finish the analysis and think "right, I need to work on X" — and then they never actually change their training. The action list needs to be pinned to the wall, added to your training plan, and reviewed at the start of every training week until the next event. Otherwise it's just an interesting exercise that changes nothing.
Prioritising limiters
If you have five things on your action list, you can't work on all five simultaneously. Joe Friel's approach is to rank limiters by impact: which one, if fixed, would make the biggest difference to your performance? Start there.
For most amateur cyclists, the priority almost always falls into one of three categories:
Pacing — If you went out too hard, no amount of extra fitness will fix the problem. Pacing is a skill. Train it.
Fuelling — If you under-ate, the fade at the end wasn't a fitness limiter, it was a nutrition limiter. You already had the engine. You just didn't put fuel in it.
Aerobic base — If you decoupled significantly, more threshold work won't help. You need more time at low intensity. This is the hardest one for time-crunched riders to accept because Zone 2 feels like you're not doing anything. But the science has finally caught up with what coaches like Friel and Kerrison have been prescribing for years — the aerobic base is the foundation everything else sits on, and if it's insufficient, nothing above it works properly.
The season-long view — tracking patterns across multiple events
A single race analysis is useful. A season of race analyses is transformative.
The real power of this framework is that it gives you consistent data points across multiple events. After three or four events analysed with the same method, patterns emerge that you simply cannot see from a single race.
The limiter tracker
Create a simple spreadsheet — or even a notebook page — with the following columns for each event:
| Metric | Event 1 | Event 2 | Event 3 | Event 4 | |---|---|---|---|---| | Event name/date | | | | | | Duration | | | | | | NP | | | | | | IF | | | | | | VI | | | | | | Aerobic decoupling % | | | | | | Carb intake (g/hr) | | | | | | Pacing (1st half NP vs 2nd half NP) | | | | | | Top limiter identified | | | | | | TSS | | | | |
When you track these across a season, the story becomes obvious. If aerobic decoupling appears in every single event, it's not a one-off — it's the thing to fix. If your VI is always high, you have a pacing habit. If your carbohydrate intake drops off in the second half of every event, it's not an accident — it's a behaviour pattern.
Dr David Dunne talks about the difference between signal and noise in training data. A single data point is noise. A repeated pattern across multiple events is signal. The season-long view is what separates signal from noise in your racing.
Comparing like with like
When you return to the same event in a subsequent year, the analysis becomes even more powerful. Same course, same climbs, same conditions (roughly). Now you can make direct comparisons.
Was your NP higher this year? Was your decoupling lower? Did you fuel better? Was your VI closer to 1.00? These are the real measures of improvement — not your finish time (which depends on conditions, group dynamics, and a dozen variables you can't control), but the execution metrics you can compare directly.
This is something Tim Kerrison understood deeply. The marginal gains philosophy at Sky wasn't about finding one big thing. It was about measuring everything consistently, finding the ten small things that were each costing 1%, and fixing them one at a time over the course of a season. The amateur version of that is this: measure the same things after every event, and fix them in order of impact.
When to repeat an event
If an event was a bad day — mechanicals, illness, terrible weather — it's noise, not signal. Analyse it for what you can learn, file it away, and don't weight it heavily in your season trends.
If an event was a good day and you executed well, that's your baseline. You now know what "good" looks like for you at that distance and that terrain. Every subsequent event can be measured against that standard.
If you're choosing which events to do next season, consider repeating at least one or two from this season. The ability to compare year-on-year with the same course is worth more than adding another new event to the calendar.
Feeding it forward
The final step is the one that closes the loop. Take your season-long limiter list, rank the items by frequency and impact, and use that to write your off-season or next-block training priorities.
If your limiters were consistently aerobic base and nutrition, your next training block should emphasise Zone 2 volume and fuelling practice. If they were consistently pacing and climbing power, your next block should emphasise negative-split long rides and power-to-weight work.
The riders who do this — who let their race data drive their training, not the other way around — are the ones who improve year after year while everyone else stays flat. It's not complicated. It's not sexy. It's not some secret training session or new supplement. It's the boring, disciplined work of looking at what actually happened and responding to it.
You've already done the hard part. You trained for months, you showed up, you raced. The file is sitting there on your computer. It contains everything you need to know about what to do next.
Forty-five minutes. A cup of coffee. The framework above. You can go and do it this week.
And the next time you line up at a start line, you won't just be hoping it goes well. You'll know exactly what you've fixed since the last one.