Do running apps actually make you faster?

Updated August 26, 2026 · By the Runked team

What an app does, and what the evidence says about it

This table is Runked's own, pairing each app capability with the published study whose measure comes closest to testing it. None of the four sources built a table like this on their own, and none of them tested app use itself as a single variable against race time.

What the evidence says about getting faster

Muniz-Pumares, Hunter, Meyler, Maunder and Smyth analyzed the 16 weeks of training behind 151,813 marathons completed by 119,452 runners, published in Sports Medicine in 2025. Average weekly volume across the full dataset was 45.1 ± 26.4 km, and the fastest finishers, those running 120 to 150 minutes, accumulated more than three times the weekly volume of slower runners in the same dataset. The correlation between volume and finish time was strong (R² ≥ 0.90), and the proportion of easy, Zone 1 running rose alongside it: more than 80% of the fastest runners used a pyramidal split, heavy on easy running with a smaller, tapering share of moderate and hard effort. The authors describe this as an association drawn from observational Strava-style data. It shows what the fastest runners' training looked like; it is not a controlled experiment, and it is not proof that copying the volume alone produces the same finish time.

The closest thing on this page to a controlled test of that intensity split comes from a smaller, older study. Muñoz, Seiler and colleagues randomized 30 recreational runners to 10 weeks of either a polarized program (about 77% low intensity, 3% moderate, 20% high) or a "between-thresholds" program weighted toward moderate intensity, published in the International Journal of Sports Physiology and Performance in 2014. This built on a 2010 descriptive review by exercise physiologist Stephen Seiler, published in the same journal, which found competitive endurance athletes training 10 to 13 sessions a week converged on roughly an 80/20 split between low- and high-intensity work (Seiler, 2010). In the 2014 trial, both groups ran faster 10-km times afterward, 5.0% for the polarized group and 3.6% for the moderate-heavy group, and that whole-group difference was not statistically significant. A post-hoc analysis of the 12 runners who most closely followed their assigned split did find a significant advantage for the polarized group (Cohen's d = 1.29, P = .038). Thirty runners over ten weeks is a narrow base to generalize from, and the significant result applies to that adherent subgroup while the whole-group comparison stayed non-significant.

Consistency is the third lever, and it shows up as a cost when it breaks. Feely, Smyth, Caulfield and Lawlor tracked training behind 509,979 marathons run by 292,323 recreational runners between 2014 and 2017, using Strava records, published in Frontiers in Sports and Active Living in 2023. Over half of runners had at least one training gap of 7 days or more before their marathon. A gap of 7 to 13 days cost about 4.25% off finish time on average, and longer gaps cost more, up to 5-8%. Timing mattered too: the paper's own summary states that disruptions occurring late in training, close to race day, were associated with a greater finish-time cost (5.2%) than similar disruptions occurring earlier in training (4.4%). This is an aggregate finding across all disruption durations; it isn't a gap-length-matched comparison, and the authors tie the difference to runners having less time to rebuild fitness before race day. The authors note they could only see runners who went on to finish a marathon, which may undercount the effect of the worst disruptions, and that the data doesn't say why a gap happened, whether injury, illness or something else.

Nielsen et al.'s 2014 study in the Journal of Orthopaedic & Sports Physical Therapy, also cited on Runked's weekly mileage guide, adds a ceiling to how fast that volume in the first study above can rise: tracking 874 novice runners for a year, it found runners who increased weekly distance by more than 30% over 2 weeks had a higher rate of specific injuries, though the difference (hazard ratio 1.59, 95% CI 0.96 to 2.66, P = .07) was not statistically significant in that exploratory sample. Volume, intensity split and consistency all assume the runner stays healthy enough to keep training, which is the boundary this study is testing.

How to apply this

• Let the app hold the plan; running the sessions stays a separate job. A scheduled week of volume, matching the structure Muniz-Pumares et al. found among faster runners, only counts once it's actually run.
• Treat feedback as a rearview mirror. Michie et al.'s self-monitoring effect came from studies where the tracked number changed the next planned session; viewing the number without acting on it is a different behavior, and it isn't the one their effect size measured.
• Respond to a flagged gap instead of dismissing it. Feely et al.'s cost was steeper for gaps late in a training block; catching one at day 7 costs less than letting it reach day 20.
• If splitting easy from hard effort, hold the split for more than a few weeks. Muñoz et al.'s significant result came from the runners who stuck closest to their assigned distribution over the full 10 weeks; the runners who drifted from it fell back into the whole-group comparison, which stayed non-significant.

Compare this month's volume and consistency to your own last month, not to someone else's training log; in Runked the rank starts everyone at Bronze and only moves on your own runs. See how the running rank works.

Common mistakes

• Treating a longer streak or a higher step count as proof of getting faster. Silva et al.'s review found apps can lift steps short-term, while the one trial in it that measured minutes of moderate-to-vigorous activity, the intensity that training plans actually target, found no significant change.
• Reading a single fast GPS split as evidence a method works. Muñoz et al.'s significant finding needed 10 weeks and close adherence to show up; one run doesn't isolate a variable the way a controlled trial does.
• Raising weekly volume by more than roughly 30% over two weeks to chase the fastest runners' numbers. Nielsen et al.'s data links jumps at that size to a higher, though not statistically significant, injury rate in novice runners.
• Ignoring a load or gap warning on the assumption the app has already accounted for it. Feely et al.'s cost estimates apply whether or not the runner expected the gap to matter.

What to do next

The volume side of this page is covered week by week in how many miles should I run a week, and the safe pace for raising that volume is in how to increase mileage safely. For the easy-to-hard split, Runked's 80/20 running guide goes through Seiler's model and Fitzgerald's version of it in more detail. And how training load, fitness and fatigue work covers the load-tracking side of the table above.

Frequently asked questions