Most people who ask how to get rid of fleas have already tried. They have bought a spot-on, applied it, watched the fleas disappear for a fortnight, and then watched them come back. The failure is almost never the product. It is a misunderstanding of what is being treated.
When you see fleas on a dog, you are looking at the smallest part of the problem. The adults on the animal are roughly five per cent of the infestation at any moment. The rest is in the carpet, in the gaps between floorboards, in the dog's bedding, under the sofa, in the car boot, and in the soil of the garden where the dog lies in the sun. Eggs fall off the host continuously and are scattered wherever the dog spends time. Larvae burrow away from light into carpet fibres and cracks. Pupae sit inside a silk cocoon that most insecticides cannot penetrate. Killing the adults on the dog does not touch any of that.
This is why flea control is measured in weeks rather than in applications, and why a plan that sounds excessive β treat the dog, treat the house, vacuum relentlessly, keep going for three months β is in fact the minimum. What follows is that plan, with the reasoning behind each part, because the parts that people skip are always the same ones.
The distribution is worth memorising because it explains every subsequent decision. In a typical household infestation, adult fleas on the host are about 5 per cent of the total population. Eggs in the environment are around 50 per cent. Larvae are about 35 per cent. Pupae are about 10 per cent.
Those figures come from the entomology literature on Ctenocephalides felis, the cat flea, which despite the name is the species responsible for the overwhelming majority of dog infestations worldwide. The numbers are approximate and vary with the stage of the infestation, but the shape is consistent: the visible problem is a rounding error next to the invisible one.
A female flea begins laying eggs within a day or two of her first blood meal, and can produce several dozen a day. The eggs are not sticky; they fall out of the coat wherever the dog happens to be, which is why the dog's favourite sleeping spot is always the epicentre. Within two to twelve days an egg hatches into a larva, a small, light-avoiding, maggot-like creature that feeds on organic debris in the carpet β including the dried blood in adult flea faeces, which is what "flea dirt" actually is.
The larva moults three times over roughly five to eighteen days, then spins a cocoon and becomes a pupa. The cocoon is the problem. It is sticky when first spun and picks up carpet fibres, dust and grit, which turns it into a camouflage shell that is mechanically tough and largely impermeable to insecticidal spray. Inside it, the developing flea can wait. Under favourable conditions it emerges in days; under unfavourable ones it enters a dormant state and can stay there for months, hatching in response to vibration, warmth and carbon dioxide β the signals of a host nearby. A vacuum cleaner passing over the carpet is, to a pupa, exactly that signal.
That single fact reorganises the whole approach. You are not trying to kill every flea in the house this afternoon. You are trying to make sure that when each one emerges β over the next three months β it lands on a dog that will kill it before it can lay.
The standard failure story runs like this: the dog is treated, the fleas on the dog die, the owner relaxes, and two weeks later the dog is scratching again. Nothing went wrong with the product. The pupae, which were never touched, hatched on schedule and found a host whose protection had lapsed or whose protection never covered the environment.
There is a second failure mode that is less obvious and more expensive: treating continuously but only the dog. An isoxazoline given every month will keep the dog comfortable indefinitely and will also, over time, reduce the household reservoir, because newly emerged fleas jump onto the dog and die before they can reproduce. But the reduction is slow and the reservoir refills from the garden and from any untreated animal in the house. In a multi-pet household, one untreated cat is enough to sustain the cycle for everyone.
The third failure is the shampoo-and-collar approach. A flea shampoo kills the fleas on the dog at the moment of bathing, which feels decisive and accomplishes almost nothing, because the environmental stages are untouched and reinfestation begins the same day. Flea collars vary enormously; some modern collars deliver a sustained dose and work reasonably, while the cheap ones are largely ineffective and occasionally irritant. Neither addresses the pupae.
None of this means products are useless. It means that the product is one of three components, and the other two β the environment and the calendar β are the ones owners underestimate.
The full cycle from egg to egg-laying adult takes roughly three to eight weeks under household conditions. The variance is not sloppiness in the literature; it is the point. Pupae in a cool, quiet corner may sit for months, while pupae in a warm, vibration-heavy living room hatch in days. Because you cannot know which is which, you cannot stop when the dog looks clear.
Twelve weeks is the working number that most veterinary guidance converges on. The reasoning: the longest plausible egg-to-adult span is around eight weeks, and pupae that hatch late in that window need a further period of protection before they can be killed. Three months covers every generation that could have been laid before treatment started, plus the stragglers.
The practical implication is that the first four weeks will look like failure. You will see fleas on a treated dog in weeks two to six, and they will be newly hatched adults that jumped on and are dying. A dog that is on a fast-acting adulticide and still has visible fleas at week three is not necessarily under-treated; it may simply be standing in a carpet that is still producing fleas. The distinction matters because the usual response β switching products β resets the clock and starts the twelve weeks again.
What actually indicates failure is a steady population past two months of consistent treatment on every animal in the house plus environmental control. At that point the differential is resistance, an untreated pet, a missed dose, or a reservoir you have not identified. Those are all solvable, but none of them is solved by buying a different brand.
The adulticide market divides into two eras, and the difference between them is large enough to be worth understanding rather than shopping by price.
The older generation is the topical spot-on: fipronil, imidacloprid, and combinations with an insect growth regulator such as S-methoprene. These work, they are inexpensive, and in some regions their reliability has been eroded by resistance, particularly in flea populations with a long history of exposure. Fipronil resistance has been documented in several countries, and the practical consequence is that a product which worked well five years ago may now underperform in a high-pressure area.
The newer generation is the isoxazoline class: fluralaner, afoxolaner, sarolaner, lotilaner. These are systemic β absorbed and distributed to the skin and tissue fluids β so a flea is killed when it bites, not when it contacts the coat. Kill speed is measured in hours, which matters more than it sounds: a flea that dies within a few hours of biting has not had time to lay viable eggs, so the reproductive cycle is interrupted on the dog as well as in the house.
Isoxazolines come as monthly chews, as three-monthly chews, and as an injectable formulation given in the clinic that provides up to a year of cover. The injectable is genuinely useful for dogs that refuse chews and for owners whose compliance is the weak link; it is also irreversible, which is a real consideration given that the class carries a regulatory warning about neurological reactions. A chew can be stopped if a dog reacts badly. An injection cannot. For a dog with a history of seizures or unexplained neurological episodes, that difference should drive the choice, and the conversation belongs with the vet who knows the history.
Two further points on product selection. Isoxazolines do not repel fleas and do not treat the environment, so a dog on one can still bring fleas into a house β they simply die before reproducing. And the class is not licensed for use in every age and weight; puppies below the labelled minimum should be treated under veterinary direction, not by splitting an adult dose.
The advice to vacuum more has been repeated so often that it reads as filler. It is not. In a flea infestation, the vacuum cleaner is doing three distinct things and only one of them is cleaning.
First, it removes. A pass over a carpet lifts eggs, larvae, some pupae, and the dried-blood debris that larvae feed on. The eggs and larvae that end up in the bag or canister are out of the cycle.
Second, and more importantly, it stimulates hatching. The vibration, warmth and airflow of a vacuum cleaner mimic a host arriving, which is precisely the cue pupae respond to. A pupa that hatches because you vacuumed is a pupa that emerges into a treated environment instead of sitting dormant for another two months. Every pass converts latent infestation into killable adults.
Third, it collects flea dirt, which is useful diagnostically. If you want to know whether the treatment is working, empty the canister onto white paper and add a drop of water. Flea faeces dissolve into a reddish-brown smear because the content is digested blood. Ordinary dirt goes grey-brown. That test costs nothing and tells you more than squinting at the dog's coat.
The frequency that matters is every three to five days during the active phase, with daily passes in the rooms where the dog sleeps. Not once a week, and definitely not a single deep clean followed by complacency. The whole point is to keep converting pupae into adults before the adulticide stops being there.
Two mechanical details get skipped. The first is the disposal: if the canister or bag stays in the cupboard for a week, you have simply moved larvae somewhere warm and quiet, and a bagless canister emptied into an open bin has the same problem. Seal it. The second is reach: eggs and larvae are in the base of carpet fibres and in the edges where carpet meets skirting board, which is where a vacuum head does least. Narrow attachments and a pass along every wall edge are worth more than another pass over the middle of the room.
An adulticide kills fleas that are already adults. An insect growth regulator prevents juveniles from becoming adults. The two do different jobs, and a plan that only has one of them has a hole in it.
The common IGRs in household products are methoprene and pyriproxyfen. Both are juvenile hormone analogues: they mimic the insect hormone that keeps a larva in a larval state, so the insect cannot complete the moult into a reproductive adult. The larva dies or develops into a sterile, non-viable form. Neither compound kills adult fleas, which is why IGR products are formulated with an adulticide rather than sold alone for active infestations.
If you are treating the dog with an isoxazoline, do you still need an environmental IGR? Strictly, not always. The dog becomes a moving trap that kills emerging fleas within hours, so the environmental reservoir is depleted at the rate that pupae hatch. In a light infestation in a single-pet household, that can be sufficient on its own.
The cases where the environmental side earns its cost are predictable: multiple pets, a carpeted house with a long-standing problem, a dog that spends its days in one room and sleeps in another, or an owner who cannot guarantee the twelve weeks of continuous treatment. In those situations, an IGR-plus-adulticide spray on soft furnishings and carpet removes the reliance on perfect compliance with the dog's medication.
What an IGR does not do is penetrate the pupal cocoon, and no product on the retail market does so reliably. This is the honest limit of chemical control indoors. The pupae come out on their own schedule; the job is to make sure that schedule lands them on a treated animal.
Soft furnishings are the parts of the environment people forget because they are not on the floor. Wash the dog's bedding weekly in hot water β 60 Β°C is the working figure for killing eggs and larvae β and dry it hot if the fabric allows. Do the same for the sofa throw, the blanket on the armchair, the towel over the crate, and anything else the dog lies on.
The car deserves its own sentence. A dog that travels gets eggs in the boot and on the back seat, and a car is a warm, dark, vibration-exposed box: close to ideal for pupae. Vacuum it out at the same time as the house, and if the infestation is long-running, treat the fabric with the same product used indoors.
The garden is the part that is genuinely difficult to control and the part most advice understates. Immature stages survive in soil, in leaf litter, under decks and in the shaded, damp spots where a dog lies. There is no realistic way to treat a lawn the way you treat a carpet. The achievable steps are unglamorous: keep grass short so the surface dries and heats, clear leaf litter and debris from the areas the dog uses, deny access to the space under the shed, and discourage foxes, stray cats and other mammals that carry fleas into the garden.
A dog that lies on the same sunny patch of lawn every afternoon is a dog that has created a small, repeatable exposure point. Moving the bed, or the patch, is free and more effective than most sprays.
Flea resistance to insecticides is real, documented, and routinely misapplied as an explanation.
The best-documented case is fipronil, where reduced susceptibility has been recorded in several countries. Permethrin and related pyrethroids have a longer history of resistance in other insect species, and the same mechanisms can appear in fleas. The isoxazolines are newer, and while the class is not immune to the problem in principle, the published evidence for field-level resistance is far thinner.
The trap is the diagnostic shortcut. An owner who sees fleas at week three concludes the product has failed, switches to another brand, and starts over. Nine times out of ten the fleas are newly hatched pupae from an untreated reservoir, and the new product will produce exactly the same result β because the product was never the problem. Switching costs money, resets the calendar, and delays the twelve-week window that would have worked.
A more useful framework: before suspecting resistance, rule out the boring causes. Was every animal in the house treated? Was the dose matched to the dog's current weight β a puppy that has grown out of its dose band is a common and quiet failure? Was the interval respected, or stretched by a few weeks? Is there an untreated reservoir, such as a cat that sleeps on the same blanket or a dog that visits a relative's house regularly? Is the environment being treated at all?
Actual product failure is characterised by a steady population despite all of those being right, over a period long enough to exclude hatching pupae. That is a conversation with a vet, and it may involve a different class of drug rather than a different brand of the same one.
For some dogs, the number of fleas is almost irrelevant. A dog with flea allergy dermatitis reacts to a compound in flea saliva, and a single bite can produce days of intense itching. These are the dogs that arrive at the clinic with the coat stripped from the base of the tail, the rump and the inner thighs, with secondary bacterial infection from scratching, and with an owner who insists there cannot be fleas because they have not seen any.
The distribution of the lesions is the diagnostic clue. Fleas prefer the lumbosacral area, the groin and the abdomen, because those are the parts a dog cannot easily groom. A dog with symmetrical hair loss and inflamed skin over the rump and tail base, with or without visible fleas, is a dog to check for flea dirt rather than to dismiss.
The management is different from ordinary flea control in one important way: the tolerance threshold is zero. A dog without the allergy can carry a small population without obvious distress. A dog with the allergy cannot, which means the twelve-week plan is not optional, environmental control is not optional, and the standard is not "fewer fleas" but "no bites". That is a higher bar than most owners expect, and it is the reason these cases relapse.
Itch control is a separate problem from flea control, and it usually needs veterinary help. Antipruritic medication, and in some cases a short course of glucocorticoids, buys the dog relief while the parasite plan takes effect. Treating the fleas without addressing the itch leaves a dog that continues to scratch damaged skin, and treating the itch without addressing the fleas guarantees a relapse. Both, or neither works.
Secondary infection is the other thing that gets missed. Broken skin from scratching colonises readily, and a dog whose skin is infected will keep itching even when the fleas are gone. If the plan is working and the dog is still uncomfortable at week four, the next question is not about fleas but about the skin.
Fleas are not only an irritant; for one common parasite they are an obligatory intermediate host.
Dipylidium caninum, the flea tapeworm, has a life cycle that runs through the flea larva. A flea larva living in carpet debris ingests a tapeworm egg. The egg develops into an infective stage inside the flea as it matures. When a dog grooms and swallows the flea β which happens constantly, because grooming is how dogs deal with itching β the tapeworm is released and matures in the gut.
The visible sign is segments: small, flat, whitish, rice-grain-like pieces around the anus, in the faeces, or on the bedding, which move when fresh and contract to a hard yellow speck when dry. Owners often describe them as "grains of rice" or "sesame seeds" and often mistake them for maggots.
The clinical significance is modest in most adult dogs and greater in puppies, where a heavy burden can cause weight loss and poor growth. The treatment is straightforward β praziquantel is the standard β but the recurrence is not, because treating the tapeworm without controlling the fleas simply restarts the cycle. A dog that has had tapeworm once and still has fleas will have it again.
The practical rule: after a confirmed flea infestation in a dog that has had fleas for more than a few weeks, ask the vet about deworming for tapeworm at the same time as starting the flea plan. It is cheap and it closes a loop that otherwise stays open.
Flea control goes wrong in a small number of cases for reasons that have nothing to do with efficacy, and these are worth stating plainly.
The first is species. Permethrin and some other pyrethroids are highly toxic to cats, and the presentation β tremors, hypersalivation, twitching, sometimes seizures β is a veterinary emergency. Dog spot-on products containing permethrin must never be applied to a cat, and a cat that grooms a recently treated dog can be exposed. In a household with both species, the product choice for the dog has to account for the cat.
The second is the isoxazoline neurological warning. Regulators have flagged a class-wide association with neurological signs in a small number of animals, and the wording matters: rare, mostly transient, but real. A dog with a history of seizures, tremors or unexplained neurological episodes should be started on this class only after a conversation with a vet who knows that history, and the owner should know what to watch for in the first days.
The third is age and weight. Every product has a labelled minimum, and the label exists because the safety data was generated at those thresholds. Splitting an adult chew for a puppy is not a smaller dose of a safe thing; it is an unlicensed dose with unknown absorption. Puppies below the labelled age are treated under veterinary direction, and often with a product licensed for the purpose.
The fourth is pregnancy and lactation. There is limited data for most products, which is not the same as evidence of harm. The decision belongs to the vet, and the default should be the product with the most established safety record rather than the newest.
The fifth is the "natural" aisle. Garlic is toxic to dogs in meaningful amounts and appears in far too many homemade flea remedies. Concentrated essential oils β tea tree, pennyroyal, eucalyptus β cause documented poisonings. Food-grade diatomaceous earth has some effect on larvae in dry carpet and no effect on the dog, and the dust is a respiratory irritant. A flea comb is the one traditional tool that does real work, because it removes fleas mechanically and tells you whether the plan is succeeding.
Indoors, yes. Central heating produces the warm, stable conditions a flea population needs, and the pupae in a carpet do not know what month it is. Outdoor exposure drops, so a dog may pick up fewer new fleas, but an established indoor cycle will continue through the coldest months. Stopping treatment in October is one of the most common reasons for a January infestation.
One visible adult usually implies a population you cannot see, given that adults are roughly five per cent of the total. It could also be a single flea picked up on a walk, which is harmless if the dog is on effective prevention. The distinguishing question is whether there is flea dirt, and whether the dog is itching. A single flea and no dirt in a dog on a twelve-month product is not a crisis. A single flea with dirt at the tail base is an infestation at an early stage.
Dog fleas will bite humans, typically around the ankles and lower legs, producing small itchy red spots in clusters. They do not live on human skin and do not infest human hair. They will live in bedding, sofas and carpet, which is the reason treating only the animal never resolves a household problem.
Plan on a minimum of eight weeks and expect twelve. The bottleneck is the pupae, which resist every product and hatch on their own schedule. Seeing a few fleas on a treated dog between weeks two and six is the plan working, not failing. If the numbers have not fallen after two months of doing all of it properly, that is when a vet or a pest control professional should be involved.
Some do, within limits. Modern collars that release a sustained dose of an active ingredient can provide months of protection, and they are useful for dogs that tolerate nothing else. The inexpensive collars sold as general repellents are largely ineffective and can cause contact irritation. No collar treats the environment, so a collar alone is the same one-legged plan as a spot-on alone.
Treat the cat, with a cat-appropriate product, at the same time and for the same duration. A single untreated cat is enough to maintain the cycle for every animal in the home, and the cat is often the source rather than the victim β Ctenocephalides felis is named for the host it prefers. This is the detail that most often explains a dog that "keeps getting fleas" despite perfect treatment, and it is worth checking before buying another product.
PawAtlas treats flea control as a scheduling problem rather than a product problem for exactly this reason: the twelve weeks are what break the cycle, and no single application, however good, can substitute for them.