How to Combine Strava and Hevy Into One Training Load Number

If you run and lift, you probably have this problem. Strava knows about the 40k you ran this week. Hevy knows you squatted heavy twice. Neither knows about the other, and so neither can answer the question you actually care about: how much did this week cost me, and can I add more next week?

This isn't a bug in either app. It's a units problem, and it's worth understanding before you try to solve it.

Why the two numbers don't add up

Strava scores your cardio using heart rate or power. Whatever the branding, it's some version of the same idea: time spent multiplied by intensity, producing a number of arbitrary units. Hevy counts tonnage — sets times reps times weight. A 5×5 back squat at 100kg is 2,500kg of work.

There is no honest way to add heart-rate-minutes to kilograms lifted. They aren't the same kind of quantity. So when people try to combine them, they usually end up doing one of three things, and all three fail in predictable ways.

Adding the raw numbers. Tonnage dwarfs everything. A moderate lifting session produces a five-figure number while a hard hour of running produces something in the low hundreds. Your "combined load" becomes a tonnage tracker that occasionally twitches when you run.

Tracking them side by side. Two charts, no answer. You can see both numbers, but you still can't tell whether a heavy squat week plus a moderate run week is more or less than a big run week with light lifting. Which is the entire question.

Ignoring the lifting. The most common approach, usually not on purpose. Your watch calls leg day "recovery" because your heart rate stayed low, so your fitness chart looks great right up until your legs disagree.

What a combined number actually requires

To add two things you need a common currency — some unit that both a squat session and a tempo run can be converted into without one of them dominating by accident.

The most established option in the sports science literature is session RPE, usually credited to Carl Foster's work in the late 1990s. It's almost insultingly simple:

Session load = RPE (1–10) × duration in minutes

Rate the whole session's difficulty from 1 to 10 once you've finished, multiply by how long it took, and you get a number in arbitrary units. A 60-minute run at RPE 6 is 360. A 45-minute lifting session at RPE 8 is 360. Same cost, different activity — and now they add.

That's it. That's the method. It works because it sidesteps the units problem entirely: instead of converting kilograms into heart rate, it converts both into perceived cost, which is the thing your recovery actually responds to.

Doing it by hand

You can run this in a spreadsheet this week. One row per session: date, activity, minutes, RPE, and load as the product of the last two. Sum by week.

After three or four weeks you'll have something more useful than either app gives you — a weekly total that includes everything. Then you can start watching the thing that matters, which isn't any single week's number but how fast it's changing.

The common way to look at that is the acute-to-chronic workload ratio: this week's load divided by your rolling four-week average. Around 1.0 means you're doing what you've been doing. Push well above roughly 1.5 and you're asking your body to absorb a week considerably bigger than anything it has been built for. Treat that as a ramp-rate signal worth noticing, not a prediction and not medical advice — the research is about population-level associations, not about you specifically.

Where the manual method falls down

Two places, and they're the reason I stopped doing it in a spreadsheet.

First, you have to remember. Every session, forever. Miss a few and your chronic baseline is quietly wrong, which makes the ratio wrong in the direction that flatters you.

Second, RPE is one number for a whole session, which is coarse. It can't tell the difference between an hour that was steady and an hour that was mostly easy with a brutal ten minutes in the middle, and it leans on your judgement on exactly the days your judgement is worst.

The more precise version

If you want to keep the two data sources you already have rather than hand-rating everything, you can compute each side properly and then fuse them:

• Cardio as heart-rate-based TRIMP, using your own calibrated zones rather than an age formula. This weights time spent by how hard your heart was actually working.
• Strength as tonnage, normalised against body weight so it stays comparable across people and across a year of getting stronger.

Then you scale the two so that sessions which genuinely cost you the same amount produce the same number, and add them.

Honest caveat: that scaling is the hard part, and nobody has a principled answer to it. A heavy squat session and a two-hour tempo ride both cost you something, but they aren't the same currency and any conversion between them is a modelling choice, not a fact. Anyone who tells you otherwise is overselling. What you can reasonably ask of a combined number is that it's consistent — that it moves when your training moves, and in roughly the right proportion.

Disclosure

I built LOAD to do exactly this automatically: it pulls cardio from Strava or Polar, lifting from Hevy or manual entry, fuses them into one weekly number and tracks the ramp rate over the combined figure. It's free, there's no card, and it backfills 12 weeks when you connect Strava so the baseline means something on day one.

But the spreadsheet method above genuinely works, and if you'd rather own your own numbers, do that instead. The important part isn't the tool. It's that you stop looking at two charts that each describe half of your week.

The methodology and citations behind LOAD's calculations are public at loadtraining.app/faq.