Almost every hatchery we visit already collects more data than it uses. Setter logs, candling counts, chick weights, customer complaints, all of it recorded somewhere and almost none of it read together. The gap is not collection. It is the missing step where a number becomes a decision that somebody is accountable for.
Key takeaways
- Hatch of set tells you what happened. Hatch of fertile plus a breakout tells you where to intervene.
- Candling at 9 to 12 days of incubation separates infertile eggs from early embryonic death reliably, which is what makes the hatch-time breakout interpretable.
- A KPI without a benchmark for that breed, that breeder age and that equipment is a number, not a signal.
- Sensor data earns its cost when it explains a breakout category, not when it fills a screen.
Start from the question, not from the dashboard
Data management in a hatchery is the systematic collection, organisation and analysis of information about the operation, but the sequence that matters runs the other way round. You decide which decision you want to make better, then you decide what would have to be true for that decision to change, and only then do you decide what to record. A hatchery that starts from the software ends up with a beautiful reporting layer over a question nobody asked.
Three decisions carry most of the value in a commercial hatchery: whether a disappointing hatch belongs to the breeder flock, to egg handling or to the machine; whether an incubation profile should be adjusted for a given flock age; and whether a customer complaint about chick quality reflects a hatchery problem or a transport and brooding problem. Every metric below exists to answer one of those three.
The one indicator that separates causes
Hatch of fertile is the indicator that separates a fertility problem from an incubation problem. Hatch of set mixes the two together, so a fall of three points tells you nothing about who should act. Hatch of fertile is calculated from first-class chicks against the eggs set minus the clears, which removes the fertility question from the denominator and leaves you looking at incubation alone.
Getting that separation right depends on when you candle. Eggs candled at 9 to 12 days of incubation are at the stage where infertile eggs and early dead embryos are easiest to tell apart, and the percentages obtained at that point can be used as the reference when the residue breakout is performed at hatch time. Candle too early and you will book early embryonic mortality as infertility, which sends a hatchery problem to the breeder farm and wastes everybody’s month.
Breakout analysis is the analytics that costs almost nothing
Residue breakout is the cheapest analytical tool a hatchery owns and the one most often skipped when the week is busy. Breaking out the hatch debris measures the impact of incubation parameters on the embryo and on what is left after incubation, and assigns each unhatched egg to a stage: infertile, early dead, mid dead, late dead, pipped not hatched, contaminated, malpositioned.
Those categories are diagnostic in a way that a hatch percentage never is. A rise in late deads and pipped-not-hatched points towards the hatcher environment and the transfer window. A rise in early deads points upstream, towards egg storage, handling and pre-warming. Contaminated eggs point at hygiene in a specific part of the flow. Three types of breakout are commonly used together, on fresh hatching eggs, on candled eggs during incubation, and at hatch time, and each answers a different part of the same question.
Benchmarks belong to a context, not to the industry
A benchmark only means something inside its context. Standards for hatchability categories are generated from local factors, breed type, source of the fertile eggs, feeding and incubation equipment, and the usual practice is to hold several standard tables that account for breeder age, which has a major effect on hatch percentage, infertility and embryonic mortality.
The operational consequence is that comparing this week’s hatch against a single company-wide target is close to meaningless when flock ages differ. Compare against the standard for that breeder age, on that line, with that equipment. Chick yield, the average chick weight at hatch divided by the average egg weight at setting, is a useful companion indicator here because it reacts to moisture loss and incubation profile rather than to fertility.
| Indicator | What it actually tells you | What it cannot tell you alone |
|---|---|---|
| Hatch of set | Commercial output of a setting. The number the customer feels. | Whether the loss was fertility or incubation. |
| Hatch of fertile | Performance of the incubation process itself, fertility removed. | At which stage of incubation the embryos were lost. |
| Residue breakout | The stage and probable origin of each loss, category by category. | Whether the cause is repeatable across flocks and machines. |
| Chick yield | Moisture loss and incubation profile, expressed as chick weight over egg weight. | Anything about fertility or about post-delivery handling. |
| Machine and environment logs | Whether conditions matched the intended profile, and for how long they did not. | Whether the deviation is what caused the observed loss. |
What sensors add, and what they do not
Sensors add resolution, and resolution is what turns a suspicion into an attributable cause. Continuous logging of temperature, humidity and carbon dioxide inside a machine tells you not only that a profile was missed but for how many hours and during which developmental window, which is exactly the information a breakout category needs in order to be explained rather than noted. Automated capture also removes the transcription errors that make manual logs argue with each other.
What sensors do not do is decide. An alarm threshold set without reference to the biology it protects generates fatigue, and a team that has learned to silence alerts is measurably worse off than one with no alerts at all. Our article on sensors and data analysis for early problem detection goes through how to set those thresholds so they stay credible.
Making the loop close in practice
The loop closes when one named person reviews the same three documents together on a fixed day each week: the hatch report, the breakout sheet, and the machine log for the settings concerned. Not three people in three departments, one person with all three in front of them. That is the single organisational change that we see convert data collection into decisions.
Two habits protect the loop once it exists. Write the interpretation down next to the numbers, so that next quarter you can tell whether the action taken worked. And keep the categories stable, because a breakout classification that quietly changes definition between operators produces trends that are pure artefact. Where a breakout keeps pointing at the same developmental window, the correction belongs to the incubation profile itself, which we cover in our guide to reducing embryonic mortality through precision incubation.
Three objections worth answering
Our hatchery is too small for analytics. Breakout analysis needs a bench, a knife and an hour, not a software licence. Sample size is the constraint for a small operation, not tooling, so aggregate several settings from the same flock before drawing a conclusion rather than reacting to one tray.
We already have a dashboard. A dashboard reports. The test is whether last quarter’s dashboard changed a decision. If you cannot name the decision, the reporting layer is running without the analysis underneath it.
The data contradicts itself. That is usually a definitions problem rather than a data problem. Two operators counting clears differently, or a chick weight taken at two different points after hatch, will produce contradictory series indefinitely. Fix the definition and the method before questioning the numbers.
The next lever, once the numbers point somewhere
Analytics stops being an accounting exercise the moment a breakout category is repeatedly attributed to a stage of incubation. What you change in the profile at that point is a separate technical decision.
See what a precision incubation profile changes stage by stage
Sources: The Poultry Site, Breakout Analyses Guide for Hatcheries; Royal Pas Reform, Hatchery Talks, egg breakout analysis; Lohmann Breeders, troubleshooting in the hatchery. Consulted August 2026. Benchmark values are site-specific and should be established with your breeding company’s management guide for the exact line.