REEF SCIENCE · GUIDE
How often should you actually ICP test?
Somewhere between "every month, always" and "once, when I panicked" there's a cadence that matches what ICP is actually good at. Here's that schedule, with the reasoning attached.
Ask this question on a forum and you'll get every answer from "monthly, non-negotiable" to "never, it's snake oil," usually in the same thread.4 Both extremes are expensive. Testing monthly on a stable tank buys you the same numbers over and over. Testing never means the first sign of a creeping contaminant is a coral dying, which is the most expensive test kit there is.
Quick disclosure so you know where I stand: I run BioForge, and our Decode service reads ICP reports for $39, so more testing is, in the dumbest financial sense, good for me. Which is exactly why I want to be precise here: most stable tanks are over-tested or under-tested, and the right cadence is probably cheaper than whatever you're doing now.
The three things worth understanding
- ICP is a slow-motion instrument. The sample spends days in the mail, and the elements it measures best, traces and contaminants, drift over weeks and months. Testing it like a daily kit wastes money.
- The schedule depends on the tank's situation, not the calendar alone. Stable tanks, new tanks, and tanks mid-problem need three different cadences.
- A test without a follow-up is half a test. Whatever you change based on a report, the retest is what confirms it. Build the follow-up into the plan from the start.
Stable tank: every 2 to 3 months
This isn't my invention. Triton, the lab that arguably created the hobby ICP category, recommends a test every 2 to 3 months even for tanks running its full method, on the logic that this is frequent enough to track trends and catch problems early.1 Quarterly is where the information rate of a stable reef roughly matches the sampling rate: trace elements like iodine or manganese deplete over weeks, contaminants from failing equipment build over weeks, and neither moves fast enough to justify monthly money.
What quarterly testing is really buying you is a trend line. One report tells you where the tank is; three or four reports a year tell you where it's going, and direction is the thing you dose against. I've made this argument before in why identical ICP results can mean two different tanks: a snapshot without history is close to unreadable.
When monthly testing earns its cost
New tanks finding a baseline
The first year of a reef is chemically noisy: rock and sand are still equilibrating, biomass is climbing, and your dosing is still being calibrated. Triton's own guidance is that new tanks and non-method users benefit from more frequent testing to establish a baseline first.1 Monthly for the first several months, then relax to quarterly once two consecutive reports look boring. Boring is the goal.
Active problems
STN, fading colors, mystery deaths: when you're diagnosing, you need data points close together, because you're trying to catch a variable in motion. Monthly, sometimes even a paired test a few weeks apart, is reasonable while the investigation is open. This is also the one time a high-value collection justifies standing monthly tests as insurance.1
After any correction: the 4-week follow-up
This is the test people skip, and it's the most valuable one on the list. If a report sent you chasing a contaminant, adjusting a trace element, or fixing salinity, the follow-up about 4 weeks later is the only thing that tells you whether the intervention worked. I walk through this logic in the order of operations: every fix ends with a retest, or it isn't finished.
The rule that matters more than frequency
Same lab. Every time. Sanjay Joshi's independent multi-lab comparison for Reef Builders showed the uncomfortable truth: send identical water to different ICP labs and the trace-level numbers come back meaningfully different.2 That doesn't make ICP useless. It makes cross-lab comparison useless. Within one lab, methodology is consistent and trends are real; across labs, you inherit every difference in their calibration and reporting.3
So a quarterly cadence with one lab beats a monthly cadence spread across three. If you're still choosing that lab, the comparison piece covers the four major options and what separates them. Pick one and marry it.
And keep the division of labor straight. Alkalinity, calcium, magnesium, nitrate, phosphate: those move on a scale of days and belong to your home kits, weekly or better. A thriving SPS tank can burn 1 to 2 dKH a day,5 which no quarterly mail-in test will ever catch. ICP is for the forty elements your home kits can't see. The two tools answer different questions, and neither replaces the other.
The budget math
As of August 2026, single hobby ICP tests run roughly $40 to $60 depending on the lab. Quarterly, that's $160 to $240 a year. One monthly-tested problem quarter adds maybe $100 more. For scale, that's one nice frag, annually, spent on knowing whether the other several thousand dollars of livestock are sitting in clean water. The cadence isn't the expensive part of this hobby.
Common questions, answered plainly
How often should I ICP test a stable reef tank?
Every 2 to 3 months. That matches Triton's own recommendation and the pace at which trace elements and contaminants actually drift. Three or four reports a year gives you a trend line, and the trend is what you dose against.
When should I test more often?
New tanks establishing a baseline, tanks with an active problem, and any tank about 4 weeks after a correction. In those windows monthly testing pays for itself; once reports turn boring, drop back to quarterly.
Does ICP replace my home test kits?
No. Fast movers like alkalinity and nutrients need weekly home testing; a mailed sample is days old on arrival. ICP covers the traces and contaminants you can't measure at home. Different tools, different questions.
Can I switch labs between tests?
You can, but your history resets. Labs disagree at trace levels, so numbers don't transfer cleanly between them. A lab switch mid-history creates phantom changes that were never in your water. One lab, one trend line.
Is quarterly ICP testing worth the money?
$160 to $240 a year, as of August 2026, to catch contaminants and trace drift before your corals report them for you. Compared to replacing lost livestock, it's one of the cheaper habits in the hobby.
Sources & further reading
WHERE BIOFORGE FITS
Make each test you pay for actually change something.
A report you don't act on is a receipt. Decode turns your ICP results into a prioritized dosing plan with a built-in retest date, so every test closes a loop instead of opening a browser tab of worry. $39, no subscription, as of August 2026.