The clamshell is one of the most widely prescribed exercises in the world for the lateral glutes, and it is also one of those surrounded by half-truths. The most common is that it is "the best for the gluteus medius". Electromyography studies say otherwise, and knowing this changes how you use it.
A note before we start, because it causes confusion: you will find it written as both clamshell and clam shell, as two words. It is the same exercise: the name comes from the clam that opens its shell, which is exactly the knee movement.
This guide starts from there: what the exercise really measures, how to perform it without wasting it, when it makes sense and when the pain you are trying to resolve does not come from the glutes at all.
What the activation numbers really say
In 2009 a group at the University of North Carolina measured the activity of the glutes of 21 healthy, physically active subjects using surface electromyography during 12 different exercises, expressing the values as a percentage of maximum voluntary contraction.
The result for the gluteus medius, in order:
- Hip abduction on the side with the leg straight: 81% of maximum contraction
- Single-leg squat: 64%
- Single-leg deadlift: 59%
- Lunges: 48%
- Jumps: 48%
- Clamshell (two variants): 40% and 38%
The clamshell sits at the bottom of the list. Side-lying hip abduction, that is, exactly the same position but with the leg straight and lifted towards the ceiling, produced roughly double the activation. The authors conclude explicitly that the best exercise for the gluteus medius, among those tested, was side-lying hip abduction.
A 2012 study of 20 healthy runners added a detail that is even less intuitive.
When an external load was applied, in the clamshell the gluteus medius worked at 32.6% of maximum contraction, but the anterior hip flexors worked at 54.2%: that is, more than the muscle the exercise is supposed to train.
The authors write that, if the primary goal is the gluteus medius, the clamshell and the externally rotated abduction are less appropriate choices than simple abduction.
Why the clamshell stays in programmes anyway
Here the second half of the picture is needed, otherwise a useful exercise gets thrown away.
In 2013 a study using fine-wire electrodes, and therefore a more precise measure than surface recording, compared 11 exercises in 20 healthy people using a different index: how much they activate the glutes compared with how much they activate the tensor fasciae latae. The tensor is a hip abductor, but it rotates the hip internally, and that is precisely the movement one often wants to avoid.
On that index the clamshell came first, with a value of 115, ahead of the lateral step (64), the single-leg bridge (59) and the quadruped exercises (50). A subsequent study of 11 healthy subjects confirmed that the clamshell, with and without a band, had the highest glute-to-tensor ratio among 13 common exercises.
Translated: the clamshell is not the exercise that loads the glutes the most, it is the one that isolates them best. These are two different qualities, and they serve different moments in rehabilitation.
What electromyography does not measure
This is the point almost nobody puts in writing, and without which the numbers above get misused.
Electromyography measures how much electrical activity a muscle generates while you perform a movement. It does not measure whether that exercise will make you feel better. These are different things: an exercise can activate a lot and produce no clinical benefit at all, and one that activates little can be exactly the right one in a phase when high load is not tolerated.
The practical limits of these studies also need considering. The samples are small, from 11 to 21 people. The participants were almost always young, healthy and asymptomatic, not people in pain.
The standard deviations are huge: in the 2009 study, side-lying abduction gave 81% with a standard deviation of 42 points, which means the variability between individuals is larger than many of the differences between exercises.
So the correct reading is not "the clamshell is not useful". It is: the clamshell is a low-load exercise, and it should be used as one.
How to perform it, step by step
- Position. Lie on one side, head resting on your arm or a cushion, hips flexed to about 60 degrees and knees bent. Feet stay together, one on top of the other.
- Pelvis. Keep it perpendicular to the floor and hold it there. This is the decisive point, and below you will see why.
- Lift. Raise the upper knee by rotating the hip outwards, like a shell opening. The pelvis does not move, the feet stay together.
- Hold. Pause at the top for 1-2 seconds, long enough to feel the glute working rather than bouncing.
- Lower. Return slowly, under control. The return phase is worth as much as the lift.
- Volume. 2-3 sets of 10-15 repetitions per side is a reasonable starting point.
The three mistakes that make it useless
The pelvis rotating backwards
This is the most common error and, unlike many technical tips going around, this one has a measurement behind it.
A 2013 study of 17 healthy volunteers compared six variants of the clamshell by changing the position of the pelvis and the hip angle. With the pelvis tilted backwards, the activation of both the gluteus maximus and the gluteus medius was significantly lower than with the pelvis neutral.
The reason is mechanical: when the pelvis rotates backwards, part of the opening of the knee no longer comes from the rotation of the hip, but from the movement of the pelvis itself. The knee rises more and the glute works less.
The practical check is simple: place a hand on the upper hip bone. If during the lift you feel the hip rolling backwards, you have already exceeded the useful range. Alternatively, lie with your back against a wall and check that the sacrum stays in contact.
A band that is too stiff
The band is not a reward, it is a variable. A band that is too resistant produces exactly the error just described: to overcome the resistance the pelvis starts to rotate, and at that point you are doing a different and worse exercise.
The rule is that the band must allow you to perform the last repetition of the set with the same quality as the first. If the last three repetitions are "snatched", the band is wrong. Start with a soft band, placed just above the knees, and move up only when the technique holds.
Forced range
Opening wider does not mean working harder. The range of the opening depends on the anatomy of your hip, and beyond a certain point the movement stops coming from the joint and starts coming from the back or the pelvis.
That individual anatomy matters is shown by a 2024 study of 30 healthy women, divided according to the angle of femoral anteversion, that is, how far the femoral neck is rotated forwards. The group with excessive anteversion had a lower glute-to-tensor fasciae latae ratio for the same exercise. The same clamshell, on the same sheet, produces different responses in different people.
The right range is the one you can achieve while keeping the pelvis still. For many people that means a modest twenty degrees, and that is perfectly fine.
The hip angle changes the result
Two studies converge on a detail almost nobody explains: how much you keep your hips flexed is not irrelevant.
In the 2013 study the activation of the gluteus medius was at its maximum with the hip flexed at 60 degrees, compared with the smaller angles tested. In the 2024 study of 30 women, the glute-to-tensor ratio was higher at 60 degrees than at 45, and at both of these angles it was higher than at 30 degrees; the upper part of the gluteus maximus also worked more at 60 and 45 degrees than at 30.
In practice: the almost-straight position, with the hips only slightly bent, is the one that gives less. Bringing the knees closer to the chest, up to about 60 degrees of hip flexion, is the more efficient choice. It is not an enormous difference, but it costs nothing.
It is worth noting that in the same study the tensor fasciae latae stayed active at low levels and was not influenced by the variants: one of the reasons why the clamshell remains a "clean" exercise.
What the clamshell is really for
Taken together, the data make the role of the exercise clear and honest.
It is a starting exercise. It is done lying down, with no weight on the hip, with no balance to manage. It is what you need when loading standing is not yet tolerated or not yet controlled.
It is useful after a period of pain. When a hip or a knee has hurt for weeks and activity has shrunk, restarting with an exercise that asks for little allows you to get moving again regularly, instead of flaring everything at the first ambitious attempt.
It is selective. If for a specific reason you want the glutes to work while limiting the contribution of the tensor fasciae latae, the fine-wire data put it at the top.
It is a good warm-up. Two sets before squats, lunges or running cost very little and help you get used to the movement of the hip.
It is not a strength exercise. With the 38-40% of maximum contraction measured in healthy subjects, strength is not built efficiently. If you stay there for months, improvements stall. This is the point at which most programmes get stuck.
An honest progression towards real load
The clamshell is the first step, not the staircase. A reasonable progression changes one variable at a time and does not skip the move to load.
- Bodyweight clamshell, paying attention to pelvis and range, until 15 slow repetitions are easy and clean.
- Clamshell with a light band, then medium, using the same quality criterion on the last repetition.
- Side-lying hip abduction with the leg straight, which in the EMG studies doubles the activation of the gluteus medius compared with the clamshell. It is the first real jump, and it is often postponed for no good reason.
- Single-leg loaded exercises: single-leg bridge, step-up, lunge, up to the single-leg squat, which in the same studies reaches 64% of maximum contraction for the gluteus medius.
- Progressive external load, adding weight to the supported exercises once pelvic control holds. If this part is unfamiliar to you, the guide on how to perform the squat correctly covers the technical points that matter.
The signal that it is time to step up is not the calendar, it is the quality: when the final repetitions are identical to the first and you no longer feel any difficulty keeping the pelvis still, the exercise has stopped giving you what it could give.
On the conditions in which the lateral glutes really are part of the problem, such as patellofemoral pain syndrome and iliotibial band syndrome, the same logic applies.
The clamshell can be the entry point, but the result comes from the loading that follows.
When the problem is not the glutes
This is the part that makes reading the article to the end worthwhile. There are situations in which continuing to do glute exercises is, at best, wasted time, and at worst a diagnostic delay.
Pain on the outer side of the hip
If the pain is pinpointed on the bony prominence on the outside of the hip, gets worse when you lie on that side, when you stand on one leg, or when you cross your legs, the most likely picture is a gluteal tendinopathy.
Gluteal tendon tendinopathy is today considered the main local cause of lateral hip pain, the condition once generically called trochanteric bursitis. It is particularly common in women after menopause.
The important thing is that here the treatment is not "strengthen more". In the most robust randomised study on this condition, 204 people with lateral hip pain lasting over three months and a diagnosis confirmed by MRI were divided into three groups: education on load management plus exercise, corticosteroid injection, or no treatment.
At eight weeks the education and exercise group reported an overall improvement in 51 of 66 cases, against 38 of 65 with the injection and 20 of 68 with no treatment, and the advantage over the injection was maintained at one year.
The detail that gets lost in summaries is that the programme was "load management plus exercise", not exercise alone.
In another study of 94 postmenopausal women, adding a targeted strengthening programme to a complete education on avoiding compression of the tendon produced no difference compared with a sham programme: both groups improved.
It is a useful reminder: in this condition, what you stop doing weighs as much as what you start doing.
A practical point that is not trivial: many people with gluteal tendinopathy simply do not tolerate lying on the affected side, which is the clamshell position. If lying on the painful side hurts, that exercise is not the right choice at that moment. The full picture is in the article dedicated to lateral hip pain.
Groin pain under load after an increase in activity
This is the point where it is worth being explicit, because it is the case in which insisting does harm.
If you have increased running, mileage, pace or sessions in recent weeks, and have developed a deep pain in the groin or in front of the hip that appears when running and eases with rest, without any trauma, that picture should be assessed before you keep training.
The reason is the femoral neck stress fracture. It is an uncommon but serious condition, typical of long-distance runners and military recruits.
It presents with insidious, non-traumatic pain in the hip and groin, which eases with rest, and it is often misinterpreted at first because X-rays are frequently normal.
MRI is the test with the best accuracy for identifying and grading it.
The detail that matters most for the reader: a negative X-ray excludes nothing. The films can remain normal for two to four weeks. And symptoms typically begin two to three weeks after an increase in the volume or intensity of training, with pain that at first appears after activity and over time persists even at rest.
If unrecognised, a femoral neck stress fracture can progress to a complete, displaced fracture, with complications such as non-union and avascular necrosis of the femoral head. Fractures on the upper-lateral side of the neck are those at highest risk.
In practical terms: in this scenario you do not keep running "to see if it goes away", and you do not add glute exercises hoping they will fix it. You stop and get assessed. Doing clamshells is not dangerous in itself, because it is a non-loaded exercise for the hip; the risk is believing it is the answer and, in the meantime, continuing to load the joint.
Night pain present even at rest
Pain that wakes you at night and is present even when you do not move at all is a different symptom from mechanical overuse pain, which typically appears with movement and improves with rest.
Night pain from compression, typical of gluteal tendinopathy, which appears when you lie on that side and changes when you put a cushion between your knees, must be distinguished from constant night pain that is independent of position.
The second kind, especially if accompanied by fever, unintended weight loss, general malaise or a cancer history, deserves a medical assessment and not an exercise programme.
This is not alarmism: it is the reason a physiotherapist, before giving you a sheet, asks questions that seem to have nothing to do with the hip.
How to fit it into the programme
- Frequency. It is a low-stress exercise: 3-5 times a week is sustainable and does not require long recovery.
- Volume. 2-3 sets of 12-20 repetitions per side, with control on every repetition.
- Position. Hips flexed towards 60 degrees, pelvis perpendicular to the floor.
- Placement. Excellent as activation before squats, lunges or running; not very sensible as the only glute exercise in a programme.
- Duration. As an isolated exercise it makes sense for a few weeks. If after a month you have not yet moved on to straight-leg abduction or supported exercises, the programme has stalled.
Useful variations
- With feet lifted. Knees on the floor and feet off the floor: increases the demand for control.
- Isometric hold. In the open position, hold for 10-15 seconds instead of returning straight down.
- Reverse clamshell. Knees together, the foot opens: works internal rotation, useful when that direction is limited.
- Standing with a band. With the band above the knees, supported on one leg, push the knee outwards. It is already a loaded exercise, so a step in the progression, not an equivalent variant.
If you are not sure what you are treating
The difference between a clamshell that helps and one that is wasted time almost never lies in the exercise itself: it lies in understanding whether the problem is really one of hip control, what phase you are in and when it is time to increase the load.
Start from the questionnaire on Ri-Hub and talk it through with a physiotherapist registered with the Albo: 20 minutes on a video call to understand where the pain comes from before choosing exercises.
In short
The clamshell is an honest exercise if it is used for what it is. In electromyography studies it activates the gluteus medius less than side-lying abduction and the single-leg squat, but it is the most glute-selective exercise compared with the tensor fasciae latae. This makes it a good starting point, a good activation and a poor end point.
Three things make it useless: the pelvis rotating backwards, a band that is too stiff and forced range. One thing makes it wasted: staying on it for months without ever moving on to load.
And if the pain is lateral and pinpointed, or groin pain that appeared after an increase in activity, or night pain even at rest, the right question is not how to do the clamshell better, but what is underneath.
Stay in touch with us
Join our dedicated Facebook group: we reply to patients, share practical advice and clear up the most common questions.
Need a consultation?
If while you do it you feel something wrong (a sharp pain, a discomfort that lingers the next day, one side behaving differently from the other), that is the moment to ask, not to keep insisting while switching videos. A physiotherapist will look at how you perform it and tell you whether it is technique, load or something to assess first: 20 minutes on a video call, free and with no credit card required.
If you do the clamshell because you have pain on the side of the hip, the condition has a name and a pathway: Trochanteric pain (trochanteric bursitis): what really works.
This exercise is part of the exercise library: eighty-three guides like this one, with performance, common errors and variations, written by our physiotherapists.
Sources
- Gluteal muscle activation during common therapeutic exercises (Distefano et al., 2009), Journal of Orthopaedic & Sports Physical Therapy
- Hip muscle activity during 3 side-lying hip-strengthening exercises in distance runners (McBeth et al., 2012), Journal of Athletic Training
- Which exercises target the gluteal muscles while minimizing activation of the tensor fascia lata? (Selkowitz et al., 2013), Journal of Orthopaedic & Sports Physical Therapy
- The influence of varying hip angle and pelvis position on muscle recruitment patterns of the hip abductor muscles during the clam exercise (Willcox & Burden, 2013), Journal of Orthopaedic & Sports Physical Therapy
- Electromyographic analysis of gluteus maximus, gluteus medius and tensor fascia latae during therapeutic exercises with and without elastic resistance (Bishop et al., 2018), International Journal of Sports Physical Therapy
- Effects of differences in femoral anteversion and hip flexion angle on hip abductor muscles activity during clam exercise in females (Mitomo et al., 2024), PLoS One
- Gluteal tendinopathy: integrating pathomechanics and clinical features in its management (Grimaldi & Fearon, 2015), Journal of Orthopaedic & Sports Physical Therapy
- Education plus exercise versus corticosteroid injection use versus a wait and see approach on global outcome and pain from gluteal tendinopathy (Mellor et al., 2018), BMJ
- Gluteal loading versus sham exercises to improve pain and dysfunction in postmenopausal women with greater trochanteric pain syndrome (Ganderton et al., 2018), Journal of Women's Health
- Femoral neck stress fractures: an updated review (Bernstein et al., 2022), Journal of the American Academy of Orthopaedic Surgeons
- Stress reaction and fractures (Bergman & Kaiser, 2025), StatPearls, 2025
Last source check: August 2026.
