"The soil looks better" is not evidence. It is a feeling, usually arriving in the third year of doing something you already believed in.
If you want to know whether your management is actually working, you need numbers you can repeat. The good news is that a small set of indicators captures most of what matters, and about half of them you can measure yourself with a spade, a coffee can and a stopwatch.
They are measurable properties that reflect how well soil functions, split across three categories: physical, chemical and biological. The major US frameworks have converged on roughly the same shortlist. The Cornell Comprehensive Assessment of Soil Health, the Soil Health Institute's measurement project and the USDA NRCS soil quality guidelines all point at the same ten or so measures.
One thing to fix in your head before you look at a single number. Most of these indicators are texture-dependent. A 3% organic matter reading is good on sand and mediocre on a silt loam. Cornell's framework uses separate scoring curves for coarse, medium and fine soils for exactly this reason. If you do not know your soil texture, the numbers mean very little.
Key Takeaway
Pick three indicators and measure them properly for five years. That beats measuring ten indicators once. Consistency of method matters more than the sophistication of the test, because almost all apparent change in soil data is sampling noise.
Four field tests are genuinely reliable in a home garden. Everything else either needs a lab or gives you a number you cannot trust.
The slake test answers whether your aggregates hold together when they get wet. Take an air-dried clod about an inch across, suspend it in a jar of water in a small wire basket, and watch. Good soil sits there. Poor soil collapses into a cloud of sediment within a minute or two. Run two jars side by side, one from your bed and one from an undisturbed edge, and the comparison does the interpreting for you.
Infiltration tells you how fast water actually enters the soil. The NRCS Soil Quality Test Kit Guide gives the standard configuration: drive a 6-inch diameter ring 2 to 3 inches into the soil, pour in 1 inch of water, and time how long it takes to disappear. Repeat in several spots because infiltration varies enormously across a few feet, and average the results.
Penetration resistance is the cheapest useful test there is. Push a screwdriver or a stiff wire flag into moist soil. If it goes 6 to 8 inches without real force, structure is decent. If it stops at 3 inches, you have a compacted layer. A proper cone penetrometer puts a number on it, and readings above 300 psi in moist soil are generally treated as root-restricting.
Earthworm counts are the simplest biological indicator. Dig a block 12 inches square and 6 to 8 inches deep, hand-sort it on a tarp and count. Fewer than 5 is low, 5 to 10 is moderate, more than 10 is good for temperate garden soil. Well-managed no-till ground can run past 20. Count after rain in spring or fall, never in a dry August, or you will measure the weather rather than the soil.
| Indicator | Where to measure | What good looks like |
| Aggregate stability (slake) | Home, jar and basket | Clod intact after several minutes |
| Infiltration | Home, 6-inch ring | 1 inch soaks in within minutes, not hours |
| Penetration resistance | Home, screwdriver or penetrometer | 6 to 8 in. easy penetration; under 300 psi |
| Earthworm count | Home, 12 in. block | More than 10 worms |
| Organic matter | Lab | Texture dependent: over 3.5% sand, over 5% loam |
| Active carbon (POXC) | Lab only | Rising trend against your own baseline |
| Soil respiration | Lab or commercial kit | Rising trend against your own baseline |
| pH and EC | Lab, or a decent meter | pH 6.0 to 7.0; EC under 2 dS/m |
Sources: USDA NRCS Soil Quality Test Kit Guide, Cornell CASH Framework Manual, NRCS bulk density indicator sheet.
Common Mistake to Avoid
Trusting a cheap home pH or NPK probe. Color-change chemical kits are roughly adequate for pH; the four-prong meters sold for houseplants are not, and their nitrogen readings are close to meaningless. If a number is going to change what you do, pay a lab for it. If it is only going to make you feel good, save your money and count worms instead.
Three things worth paying for. Organic matter percentage, because it is the integrating number that everything else tracks against. Active carbon, or POXC, which measures the fraction of organic matter that microbes can actually eat. And soil respiration, which measures how much biology is currently at work.
Active carbon is the one most people have not heard of and the one that pays off fastest. Total organic matter barely moves in three years. Active carbon does, which makes it an early-warning signal that your management is working before the headline number confirms it.
Cornell's Soil Health Laboratory prices a full panel at $100 a sample, with active carbon and wet aggregate stability at $30 each if you want them individually. Your state extension lab is cheaper for a basic fertility panel, often under $30, but most do not offer the biological measures. Both are worth having: fertility to fix a deficiency now, soil health to tell you whether the system is heading in the right direction.
Why This Works: Continuous Living Roots
The practices that move these numbers all do one thing: they keep something growing. Root exudates are the primary carbon input feeding the microbial community, and a bed that sits bare for four months has switched that supply off. This is the mechanism behind cover cropping, and it is the same reason successional systems stack plants across time. Syntropic agriculture treats bare soil as the problem to be designed out rather than a phase to be tolerated.
Get Our Free Companion Planting Chart
Join 10,000+ gardeners getting weekly tips on what to plant together, soil health, and permaculture techniques.
Send Me the ChartSlower than the internet suggests, and at very different speeds depending on the indicator.
University of Minnesota Extension puts a realistic organic matter goal at about 0.1% a year under soil-building practices. It also gives the detection timeline, which is the part people skip: soils starting at 0 to 2% organic matter may show a change within three to five years, while soils already at 4 to 7% can take seven to ten years or more. A 20-year record from the University of Washington student farm under regenerative management found a weak increase of around 0.5% per year, which is at the optimistic end of what is published.
Active carbon and respiration typically respond within one to three years. Total organic matter, bulk density and mean aggregate stability usually need five to ten years of consistent management before the trend rises above natural variation.
Fix your sampling spots and mark them
Choose three to five locations per bed or block and record them. Composite the cores into one sample. Moving your sampling spots between years is the single commonest way gardeners fool themselves.
Fix the depth
Zero to 6 inches for gardens, or 0 to 8 for cropped ground. Write it on the bag. A sample taken to 4 inches will read higher in organic matter than one taken to 8 from the same bed.
Fix the time of year
Same month every time, ideally fall after harvest. Respiration and worm counts swing hard with soil temperature and moisture, so a spring versus fall comparison tells you about the season, not the management.
Retest on a two to three year cycle
Annual testing costs money and mostly measures noise. Two to three years is the shortest interval at which a real trend can separate itself from sampling variation.
Keep an unimproved control if you can
A strip of ground you leave alone, or a neighboring lawn, gives you a comparison that absorbs the year-to-year weather. Without one, a dry year can look like management failure.
Key Takeaway
The honest reading of the evidence: cover crops, reduced disturbance and organic matter additions reliably move multiple indicators at once. The measured data for agroforestry and successional systems specifically is thinner and more geographically patchy than the data for cover cropping and no-till, which is worth knowing before you cite it at anyone.
Function, not composition. Healthy soil takes in water quickly, holds it, releases nutrients on a schedule plants can use, resists erosion and supports a large and varied population of organisms. That set of behaviors emerges from structure and organic matter rather than from a particular nutrient level, which is why a soil can test adequate on nitrogen, phosphorus and potassium and still grow poor crops. The indicators in this guide are chosen because each one measures a different piece of that functioning, and no single number captures it.
Four tests cover most of it. The slake test for aggregate stability, a 6-inch ring for infiltration, a screwdriver pushed into moist soil for compaction, and a 12-inch soil block hand-sorted for earthworms. All four cost almost nothing and all four are more informative than a hardware store meter. Their value comes from repetition: run them at the same time of year, in the same spots, and you build a record. Run them once and you have an anecdote.
A basic fertility panel from a state extension lab typically runs under $30 and reports pH, nutrients and usually organic matter. A full soil health panel including biological measures is considerably more: Cornell's laboratory charges $100 per sample, with active carbon and wet aggregate stability available separately at $30 each. For most home gardeners, the sensible pattern is a basic panel to catch a nutrient or pH problem, then a full panel every few years if you are actively trying to rebuild a degraded soil.
Five practices carry nearly all the measured effect: keep living roots in the ground as much of the year as possible, disturb the soil as little as possible, keep the surface covered, add organic matter, and grow a diverse mix rather than the same thing repeatedly. In garden terms that means cover crops in the off-season, compost applied to the surface, mulch on any bare ground, and giving up the annual dig. The no-till evidence is the best documented of the five.
It depends entirely on texture, which is why a single figure is not useful. Cornell's framework scores coarse sandy soils as low below roughly 1.5 to 2%, moderate at 2 to 3.5% and high above 3.5 to 4%. Medium-textured loams and silt loams score low below 2 to 3%, moderate at 3 to 5% and high above about 5%. Fine clays hold more again. Your own baseline matters more than any benchmark: a sandy soil moving from 1.8% to 2.4% has improved considerably, even though the second number still looks unimpressive on a chart.
Ready to Grow Smarter?
Get our free beginner's guide to permaculture gardening, 12 pages of practical tips you can use this weekend.
Download the Free Guide