Should cat and dog campaigns really be separate, or can one campaign handle both?

Separate them. Species is not a soft preference the delivery system can optimise around: a dog food ad has no purchase probability for a cat owner, so mixing the two means part of every impression pool is unreachable. Keep separate audiences with reciprocal exclusions, and put households that have bought for both animals into both audiences rather than into a third mixed one.

How do I work out when a customer will run out of food?

Divide the pack weight by the daily ration printed in the feeding guide. A 12 kg bag at 300 g a day is 12,000 divided by 300, or 40 days of supply. Do this per SKU, then check the result against the median number of days between first and second orders of that SKU in your own data, and trust the observed figure when you have enough orders to read it.

Why is a single 30 day purchaser exclusion a problem for a pet food account?

Because different packs empty on different dates. If one SKU lasts 40 days and another lasts 30, a blanket 30 day rule releases the 40 day buyer while they still have 10 days of food left, and keeps the 30 day buyer excluded through the entire window when they were ready to reorder. Set the exclusion length from each SKU's days of supply instead.

Where should emotional creative be used and where should it be avoided?

Use it at the top of the funnel, where the job is to stop the scroll and the emotional bond with the animal is your strongest asset. Avoid it in reorder retargeting, where the buyer already knows the brand and needs the pack, the size, the price and a fast path to checkout. The common mistake is promoting a high performing cold creative into retargeting, where attention was never the bottleneck.

Do toys, beds and other accessories fit the same consumption cycle logic?

No. Hard goods have no depletion date, so their trigger is damage, the pet outgrowing a size, or a life stage change rather than an empty bag. The practical use for them is the suppression window: during the weeks you should not be selling that customer more food, accessories are the natural cross sell. Keep their return target separate from food, since the margins and repeat behaviour differ.

Meta Ads for Pet Brands: Playbook

Two structural facts decide whether meta ads for pet brands work, and neither of them is creative. A cat owner and a dog owner are different audiences who happen to shop in the same store. And pet food is a consumable, so the date of the next purchase is something you can calculate rather than guess at. Ignore either one and you spend the budget showing a 12 kg bag of dog food to somebody with a cat, on a day when their cupboard is already full.

Why meta ads for pet brands start with the species split

Species is not a preference dimension like colour or style. It is a hard boundary. A dog food ad shown to a cat owner is not merely less relevant, it has no purchase probability at all, no matter how good the creative is. That makes species the first cut in the account structure, above interest, above lookalike source, above everything.

In practice, purchase and content view events have to carry a species value, or you need to derive one from the product type in your catalogue. Without that field you cannot build a cat buyer audience, you cannot exclude cat buyers from dog campaigns, and every downstream idea here is unavailable. Whatever platform runs the audience layer, the field has to exist in the data first.

Two details are worth getting right early. Households with both animals belong in both audiences, not in a mixed bucket that receives neither campaign's creative. And resist splitting further than your volume supports: small breed and large breed food are genuinely different products, but a split that leaves each ad set with a handful of weekly conversions trades relevance for instability. How far to divide an audience is covered in audience segmentation strategies.

Days of supply is the number the rest of the account hangs on

For every food SKU you advertise, work out how long one pack lasts a single animal. The inputs are printed on your own packaging, in the feeding guide, so nothing here has to be borrowed from outside data.

Suppose your best selling dog SKU is a 12 kg bag and the feeding guide gives 300 g a day for the weight band that buys it most. That is 12,000 g divided by 300 g, which is 40 days of supply. Now take a 1.5 kg cat bag with a 50 g daily ration: 1,500 divided by 50 is 30 days. The dog pack lasts one third longer than the cat pack, so the two cannot share a retargeting schedule even though they sit in the same catalogue.

Multi-pet households move the number a long way. Two dogs eating 300 g each consume 600 g a day, so the same 12 kg bag empties in 12,000 divided by 600, which is 20 days, exactly half the single animal figure. You will not know pet count for most buyers, so treat the single animal number as the default and let observed data correct it.

That correction matters more than the calculation. Once you have enough repeat orders, pull the median gap in days between a customer's first and second purchase of the same SKU. If the pack arithmetic says 40 days and your order history says 34, use 34. The feeding guide describes intended use; your orders describe actual use.

Turning days of supply into a retargeting calendar

With a number per SKU you can derive retargeting windows instead of guessing them. Using the 40 day dog bag, the calendar has three phases:

Days 0 to 31, suppression. The buyer has food. Exclude them from prospecting and from every reorder message for that SKU. This is not dead time: it is the natural place to sell everything that is not food.
Days 32 to 40, the reorder window. The final fifth of the supply, which for a 40 day pack is the last 8 days. This is the only stretch where a reorder ad is timely rather than irritating.
Day 54 onward, win back. Two weeks past depletion, the customer has bought food somewhere and it was not from you. The message changes, because you are asking someone to switch back rather than to repeat.

Run the same arithmetic on the 30 day cat pack and the phases land on days 0 to 23, then days 24 to 30 for a final fifth of six days, then day 44 for win back. Same method, different dates.

This is where the most common mistake in the category becomes visible. Plenty of accounts run one blanket exclusion, something like purchasers in the last 30 days, across everything. Applied to the two SKUs above it fails in both directions at once. The cat buyer's reorder window opens on day 24 and closes on day 30, entirely inside the exclusion, so you never reach them during the only week that mattered. The dog buyer re-enters the pool on day 30 with 10 days of food still in the cupboard. One number cannot serve two depletion dates. Custom audience retention settings are the mechanism here, so read Meta's current documentation before committing to a retention length.

Emotional creative earns attention, functional creative earns the reorder

Pet brands have an unusual creative advantage: the product is bought for a family member the buyer is genuinely attached to. At the top of the funnel, where you compete for attention against everything else in the feed, an animal's face stops the scroll in a way an ingredient panel never will.

The failure mode is promoting that same asset down the funnel because it tested well on click through rate. Someone eight days from running out of food does not need reminding that they love their dog. They need the pack shot, the weight, the price, the delivery time and the shortest route back to the product page. At the bottom of the funnel, attention is not the constraint. Friction is.

So the split runs roughly like this. Cold audiences get the emotional hook with a functional payload behind it: the first second is the animal, the rest is what the product does. Mid funnel is where sourcing, formulation and whatever claims your packaging and regulatory position genuinely support belong. Reorder audiences get product first creative with almost no storytelling. Because those audiences are small and see the same handful of ads repeatedly, that is where creative fatigue shows up first, so plan a rotation for the bottom of the funnel before you plan one for the top.

Accessories and hard goods do not obey the cycle

Everything above applies to food and to other consumables such as litter and treats, where the same weight divided by usage calculation works. Hard goods do not behave that way. A bed, a carrier, a bowl or a harness has no depletion date. Replacement is triggered by damage, by a puppy or kitten outgrowing a size, or by a change in life stage. The clock is the pet's age, not the cupboard.

Suppression leaves you 31 days in which you should not be selling that customer more food, and hard goods are what fills the gap: treat accessories as a cross sell layer running inside the suppression window rather than a separate campaign competing for the same impressions. What you should not do is hold food and hard goods to a single return target, because the margins and the repeat behaviour differ.

A build order that gets this live

The sequence matters: each step is useless without the one before it.

Get species onto the event, as a parameter on purchase and content view or derived from the catalogue product type. If that is not possible, fix it before spending anything else.
Build a sheet of your food SKUs with pack weight, daily ration and the resulting days of supply. Most catalogues cluster into two or three groups rather than needing a window per SKU.
Compare each computed figure against the observed median gap between first and second orders, and replace it wherever the observed number is reliable.
Build suppression, reorder and win back audiences per species per group, with reciprocal species exclusions.
Assign creative by funnel stage rather than by asset performance, and set a rotation schedule for reorder audiences.

Judge the result on second orders rather than first order return. In a consumable category the acquisition cost of the first bag only becomes readable once you know how many bags follow it, and a CAC calculation that ignores the reorder is measuring the wrong thing. The pet food version of that work starts here, with a number you can read off the back of your own packaging.