Original Finding
Three Feeders Reach 93% of Feeder Birds, and a Tube Is Not One of Them
Treating "which feeders should I buy" as a set-cover problem over our 73 feeder-using species gives a clear order: platform first, suet cage second, nectar feeder third. The tube feeder that most beginners buy first comes fourth, and adds 4 species.
The short answer
- A platform or tray alone reaches 46 of 73 feeder-using species (63%), more than any other single feeder type.
- Adding a suet cage brings the total to 61 (84%). Adding a nectar feeder brings it to 68 (93%).
- A tube feeder ranks fourth and adds 4 species, because almost everything a tube serves is already served by the platform.
- This measures breadth of species reached, not volume of birds or your enjoyment of watching them. Read the limits section before you sell your tube feeders.
Method at a glance
- Question
- What is the smallest set of feeders that reaches the most species?
- Data
- Feeder bird trait matrix, 73 feeder-using species
- Method
- Greedy set cover over normalized feeder-type tags
- Measure
- Distinct species reached, unweighted by abundance or visit frequency
- Reproducible
- Yes,
research/analysis.pyagainst the published CSV - Type
- Analysis of compiled data, not field observation
The question, stated properly#
"What feeder should I buy?" is usually answered one feeder at a time, which is the wrong shape for the question. Feeders overlap heavily: two feeders that each reach 40 species might reach 45 between them or 75, and which one it is decides whether the second purchase was worth making.
The right shape is set cover. Given each species and the feeder types it will use, find the smallest set of feeders that reaches the most species, and get the order of purchase out of it. We ran the standard greedy version: repeatedly add whichever feeder brings in the most species not already covered. Greedy is not guaranteed optimal for set cover in general, but it answers the practical question directly, which is what to buy first, second and third.
The result#
| Buy | Feeder | New species added | Cumulative | Coverage |
|---|---|---|---|---|
| 1 | Platform or tray | +46 | 46 | 63% |
| 2 | Suet cage | +15 | 61 | 84% |
| 3 | Nectar feeder | +7 | 68 | 93% |
| 4 | Tube feeder | +4 | 72 | 99% |
| 5 | Dish, cup or jelly feeder | +1 | 73 | 100% |
Four feeders reach 99% of the feeder-using species in the matrix. Three reach 93%. The curve is steep and then flat, which is the useful part: the difference between owning nothing and owning one platform is 46 species, and the difference between owning three feeders and owning five is 5.
Each feeder on its own#
Set cover rewards a feeder for what it adds, which penalizes latecomers with heavy overlap. Here is the same data without that penalty: how many species each feeder type reaches on its own.
| Feeder type | Species that use it | Share of feeder birds |
|---|---|---|
| Platform or tray | 46 | 63% |
| Suet cage | 25 | 34% |
| Ground or scattered | 25 | 34% |
| Hopper feeder | 19 | 26% |
| Tube feeder | 19 | 26% |
| Peanut feeder | 11 | 15% |
| Nectar feeder | 9 | 12% |
| Dish, cup or jelly feeder | 9 | 12% |
| Nyjer or finch feeder | 3 | 4% |
Note the tube's standalone number against its set-cover contribution of 4. It is not that tubes reach few species. It is that a platform has already reached nearly all of them.
Why the tube ranks fourth#
Two reasons, and they compound. The first is the mass ceiling: the tube list ends at about 44 g, so every heavier species is unreachable by design. The second is that the species below the ceiling, the chickadees, titmice, finches and nuthatches, will nearly all use a platform too. So the tube's exclusive audience is small: birds light enough to be tube-compatible and somehow unwilling to use a flat surface.
A suet cage ranks second for the opposite reason. It reaches an audience almost disjoint from the platform's: woodpeckers, creepers, and the insect-eaters that largely ignore loose seed. Low overlap is what makes a second purchase worth making.
The general principle, which survives past this particular dataset: buy your second feeder for what it reaches that your first one does not. Most feeder advice ranks feeders individually, which systematically over-recommends the second tube.
What this does not show#
Species reached is not birds seen. Coverage counts distinct species, so a feeder serving one rare species scores the same as one serving forty House Finches. If your goal is a busy window rather than a long list, weight this differently.
Platforms have real costs this analysis ignores entirely. They soak seed in rain, they feed squirrels and large flocking birds happily, and an open tray raises disease-transmission risk because droppings land in the food. Our cleaning guide and squirrel guide both exist largely because of platforms. A tube with a weather guard is the lower-maintenance, lower-risk object, and this analysis gives it no credit for that.
Coverage is binary and unweighted. "Uses this feeder" does not distinguish a staple from an occasional visit. Visit-frequency data would change the ordering and is exactly what the Feeder Log is gathering.
Regional mix matters more than this table admits. Nectar feeders rank third partly because hummingbirds are a large, distinctive block that nothing else serves. In a region with one hummingbird species for four months a year, that ranking moves.
What to do with it#
Starting from nothing and wanting range: a platform or a hopper with a wide ledge, then a suet cage, then a nectar feeder in season. That is three purchases and, on this data, 93% of the feeder-using species in the matrix.
Starting from a tube, which is where most people start: the next purchase that adds the most is a suet cage, and the cheapest upgrade of any kind is a tray fitted under the tube you already own, which converts it into a platform for the price of a coffee.
Then read the counterweight to all of this: cleaning, because a platform you do not clean is worse for birds than a tube you do.