Home How Greenhouse Match Scores Work
How Greenhouse Match Scores Work
The Greenhouse Finder uses a documented, deterministic fit score. It is designed to explain why a particular greenhouse fits a particular set of inputs—not to manufacture a universal product rating.
A 90/100 result means a cataloged configuration aligns strongly with the entered buyer requirements under the current scoring rules. It does not mean the product is “90% good,” 90% safe, or better built than every lower-scoring greenhouse.
1. Physical fit comes before points
The Finder first checks whether each product's documented exterior width and length fit inside the visitor's maximum site width and length. It checks both orientations. If the greenhouse does not fit either way, it is excluded before any score is calculated.
This is intentionally stricter than giving a large greenhouse a low “space score.” A product that physically cannot fit the entered site is not a useful recommendation, no matter how attractive its glazing or other features might be.
If a visitor also chooses a target size class, the Finder applies a second capacity guardrail. A product below 65% of the target floor area is excluded. That prevents a very small greenhouse from ranking as a substitute for a size recommendation that was based on substantially more growing capacity.
2. Current 100-point weighting
For products that clear the footprint rules, the Finder combines eight buyer-fit categories. The maximum values are transparent:
| Category | Maximum points | What it evaluates |
|---|---|---|
| Space / target size | 35 | How efficiently actual floor area fits the site and, when selected, the intended size class. |
| Growing goal | 10 | Season-extension, propagation, general-season, overwintering or warm year-round intent. |
| Heating plan | 5 | Whether the relative glazing/envelope class is sensible for the selected heating intent. |
| Climate priorities | 20 | Presence of manufacturer-published snow/wind information and/or documented ventilation for overheating concerns. |
| Crop / primary use | 10 | Peak-height implications for tall crops or general fit for trays, containers and mixed growing. |
| Glazing preference | 10 | Whether the documented glazing matches polycarbonate, glass or film preference. |
| Structure preference | 5 | Residential rigid kit, heavy-duty, glass or high-tunnel shopping preference. |
| Assembly tolerance | 5 | Broad alignment with the cataloged moderate/advanced assembly class. |
| Total | 100 | Buyer fit for the entered scenario. |
If the visitor does not identify snow, wind or overheating as a special concern, the climate category is neutral at its full 20 points. A product is not penalized simply because the visitor chose not to make climate documentation a differentiating factor.
3. How the 35 space points work
Space receives the largest weight because a greenhouse that does not suit the physical project is a poor recommendation regardless of other features. If no nominal target is selected, the Finder compares product floor area with the available site area after confirming the actual dimensions fit. A greenhouse that uses a meaningful portion of the site efficiently can earn more space points than one that leaves most of the available footprint unused.
If a target class is selected, the logic changes from “use the site” to “meet the intended capacity.” Products around 100%–135% of the target area receive the strongest space score; somewhat larger products can still score well if they fit. Slightly smaller products lose points, while configurations far below the intended capacity are excluded.
This target is deliberately based on floor area rather than a literal nominal label. An actual greenhouse sold as 8×12 may not be exactly 96 square feet, and a nominal 10×12 may have different exterior dimensions from another 10×12.
4. Relative envelope classes are not invented R-values
Heating-related scoring needs some way to distinguish a basic clear or film enclosure from more insulation-oriented multiwall glazing. GreenhouseMatch therefore stores an internal relative envelope class for matching: basic, moderate, strong or very strong.
That class is a comparison aid derived from documented glazing construction. It is not presented as a manufacturer R-value, U-factor or tested whole-building thermal performance. Where an exact thermal value is not documented, the Finder does not invent one. The envelope class simply prevents a heated year-round plan from receiving exactly the same heating-fit points as an unheated plan when the documented glazing constructions differ substantially.
5. Weather documentation earns evidence points—not a safety declaration
When a visitor selects heavy snow or high wind as a concern, a configuration with stored manufacturer-published numeric information earns stronger climate-documentation points. A product with no stored numeric claim receives only partial credit and the result calls out the missing information.
This is an evidence filter, not a local design calculator. The Finder does not know the required design load for every address, does not certify anchoring and does not convert a manufacturer marketing statement into code compliance. Published figures may depend on configuration, anchoring, assembly, maintenance, snow removal or other conditions.
If current sources conflict, the conflict belongs in the catalog and result notes. GreenhouseMatch's specification-verification policy explicitly favors showing disagreement over silently choosing whichever value makes the product look stronger.
6. Why price is not currently a Finder input
The catalog can retain dated observed-price fields where an exact configuration has a defensible price source, but those fields are not currently used by the Finder. Price coverage is uneven across brands: some manufacturers or distributors publish direct configuration prices, while others primarily publish specifications and send shoppers to retailers whose prices, promotions, shipping treatment and regional availability vary.
Using that incomplete dataset as a maximum-budget filter would create false confidence. A product with no comparable verified price cannot responsibly be treated as either under or over a visitor's budget. For that reason the Finder currently leaves price out of eligibility, scoring, visible result specifications and Data Confidence. Price can still be used selectively on exact model/comparison pages or entered directly by the visitor in the Project Cost Calculator.
7. Data confidence is separate from match score
A product can fit the visitor's needs well while still having incomplete documentation. To keep those concepts separate, each result includes a Data confidence label.
The current confidence check awards completeness points for these fields:
- actual width and length;
- published peak height;
- glazing material and construction;
- some stored ventilation information;
- manufacturer snow or wind information;
- at least one primary manufacturer/distributor source.
Actual footprint receives extra weight in this completeness check because it is fundamental to the Finder's hard fit rule. A record with 6–7 confidence points is labeled High; 4–5 is Medium; below 4 is Limited.
Data confidence is not a judgment of product quality. “Limited” can simply mean the manufacturer has not published enough of the fields the Finder uses, or that the initial catalog has not yet verified those fields.
8. Result labels are thresholds, not editorial awards
The uniquely highest-scoring result when its score is at least 78 and data confidence is not Limited. If the top products tie, none receives this label.
A result scoring at least 74 after the hard fit checks.
A result from 60–73 where the product may fit but important compromises or gaps deserve attention.
A displayed result below 60. Products below the Finder's display threshold are not shown.
The exact number should never be read without the “Why it matched” and “Tradeoffs / verify” lists beside it. Those explanations are more important than a one-number ranking. When products tie on fit score, the Finder orders them deterministically by data-confidence level, then by floor area and product ID; that tie-break ordering does not add hidden score points.
9. Affiliate economics are outside the scoring function
The scoring code does not receive commission rate, cookie duration, affiliate approval status or merchant payout as inputs. A product's merchant route is attached to the result only after the product has been evaluated from its fit and documentation fields.
That separation matters because GreenhouseMatch may have different commercial relationships for a general greenhouse kit, a heavy-duty greenhouse, a high tunnel and a premium glass greenhouse. Those relationships can determine where a “Browse” link goes; they cannot turn a poor physical fit into a better match.
The Finder is allowed to rank a product regardless of whether a current comparable retail price is stored or whether its merchant relationship is weaker than another product. The result should represent the project first and monetization second.
10. Why the score can change over time
A deterministic score can still change when the underlying evidence changes. A manufacturer may revise dimensions, introduce a new generation, change included vents, publish additional structural information or discontinue a configuration. Dated price observations retained for other tools may also be updated, but they do not alter Finder ranking.
That is why each catalog record carries verification metadata and primary source links. Changes should be traceable to product data or scoring rules rather than unexplained editorial preference. If the scoring model itself changes materially, this methodology page and the Finder should be updated together.
11. What we would rather leave unknown
GreenhouseMatch intentionally leaves a field missing when the available source does not justify a confident value. Examples include an unverified vent-opening area, a weather number copied only by a dealer, a warranty that varies by component, or a price that cannot be tied to the exact configuration.
A blank field is less convenient than a complete-looking comparison table, but it is more useful than a precise number that cannot be defended. That principle applies equally to the Finder, future comparison tools, heating/ventilation calculations and product pages.
Frequently asked questions
Is a Greenhouse Match score a product quality rating?
No. It is a scenario-specific fit ranking. The same greenhouse can score differently for a small unheated seed-starting project and a larger heated overwintering project.
What is data confidence?
It is a separate measure of how complete the decision-relevant catalog record is. A high fit score with Medium confidence means the product aligns well with the known inputs, but some important evidence is still missing.
Can affiliate commission improve a product's score?
No. Commission and merchant economics are not scoring inputs.