What to Automate First
Automation trades human error for mechanical error. That trade is strongly favorable for some tasks and clearly unfavourable for others, and the difference is whether the task rewards consistency or judgement.
The test
Automate tasks where doing the same thing every day is the goal and a human is unreliable at it. Do not automate tasks where the point is noticing something, because a machine that performs the task also removes your reason to look at the tank.
| Task | Automate? | Why |
|---|---|---|
| Photoperiod | First | Pure consistency, zero judgement, and erratic lighting is a leading cause of algae. Cheapest device on the site |
| Top-off | Second | Evaporation shifts salinity daily in a marine tank. A human doing this by eye is worse than a float valve |
| Temperature | Already automatic — add a backup | The heater automates it; the value you can add is an independent cutoff, not more automation |
| Dosing | Once the demand is known | Consistency matters, but only after you have measured what the tank actually consumes |
| Feeding | Reluctantly | Feeding is when most people look closely at their fish. Automating it removes the daily inspection |
| Water changes | Late, if ever | Highest consequence if it goes wrong, and it is when you notice detritus, equipment and behavior |
Why dosing comes after measurement
A dosing pump delivers what you tell it, precisely, forever. If what you told it is wrong, it delivers that wrongly forever too — and alkalinity dosed at the wrong rate can take a reef down in hours.
The sequence that works is: test, establish what the tank consumes over a week, dose manually to confirm, then automate the confirmed amount. Automating first means you have built a machine to repeat a guess.
Dose small and often. Whatever the daily amount, splitting it across many small doses is gentler on parameters than one large one — and it means a runaway failure delivers less before you catch it. The sizing is itself a safety measure.
The hidden cost: you stop looking
The strongest argument against automating everything is not mechanical. A well-automated tank runs without you, and a tank you have no reason to visit is a tank whose problems get found late.
Feeding is the clearest case. It is the one daily task that puts you in front of the glass watching the animals behave — which is how illness, aggression and equipment problems are actually noticed. A feeder is genuinely useful for absences and a real loss as a permanent arrangement.
If you do automate broadly, replace the removed contact deliberately: a fixed time each week to look properly, test, and check equipment you no longer touch.
No page here promises a livestock outcome. We keep no tank and test nothing. Every automated device is a mechanism operating unsupervised beside animals that cannot leave, and none of them is a substitute for observation.
Automate what fails quietly
The equipment worth buying first is whatever prevents a slow, silent drift — temperature, top-off, dosing. Those are the parameters that move gradually enough to go unnoticed and far enough to matter, and every one of them has ended a tank. Feeding and lighting are conveniences by comparison: useful, and not what loses livestock. Ordering purchases by what fails quietly rather than by what is most visible is the single most useful habit in equipping a tank.
Detection before logging
A device that tells you something is wrong is worth more than one that records what happened. A temperature alarm that reaches a phone earns its place before a monitor that graphs the last month, and a leak sensor that cuts a pump earns its place before either. The graph is useful once the tank is stable; the alarm is what gets it there. Buying in the other order is common and expensive.
What automation does not cover
No amount of equipment notices a fish in distress, a filter beginning to block, or food that has stopped being eaten. Automation removes the jobs that are regular and mechanical, which is exactly the set that does not include looking at the tank. A week away needs a feeder and a top-off and someone willing to glance at it once, and the third of those is the one that catches what the other two cannot.
Top-off before dosing
Evaporation concentrates everything left behind, so a tank losing water is a tank whose salinity and parameters drift daily. An auto top-off holds the volume steady and makes every other reading meaningful, which is why it comes before dosing rather than after: dosing into a tank of changing volume is adding precision to a moving target. It is also the automation most likely to cause a flood, which is why it belongs with a leak sensor.
Buy for the week you are away
The honest test of any automation here is what happens during a week nobody is present. A feeder that clogs, a top-off with no leak sensor, a controller whose alarm sounds in an empty room — each is adequate day to day and inadequate for the case it was bought for. Setting the equipment up and then leaving the tank alone for a week while still at home is the only way to find out which is which.
What to skip
Equipment that produces a number nobody acts on is the recurring waste in this hobby. A monitor logging a parameter that has never moved, a second sensor for a failure that has never occurred, an app for a device with a display already on it. The test is simple and unforgiving: if the reading would not change what you do next, the money is better spent on redundancy for something that would.
The one thing worth buying before any of it
A spare heater sitting in a cupboard costs very little and turns a failure from an emergency into an inconvenience. The same applies to a spare impeller, a spare airline and a length of tubing. None of it is automation and all of it addresses the same problem automation is bought for, which is what happens when something stops working at the least convenient moment. Equipment that removes the panic is worth more than equipment that removes the chore.
Common questions
What is the single best thing to automate?
The photoperiod. It is the cheapest, it has no failure mode worse than lights not switching, and it fixes a genuine and common problem.
Should I automate water changes?
It can be done well and it removes one of the best opportunities to inspect the tank. If you do, schedule inspection separately and deliberately.
Is a feeder a bad idea?
Not at all for absences. As a permanent replacement for daily feeding it costs you the observation that daily feeding provides.
Do you recommend specific automation setups?
No. This is an ordering of tasks, not of products.