Commercial beekeepers make difficult varroa management decisions throughout the season: when to treat, which treatments to use, how many applications to make, and whether the results justify the time and expense.

Experience is essential, but evaluating and improving protocols across thousands or tens of thousands of colonies requires something more: consistent monitoring, reliable treatment records, and a way to see how the two relate.

Until recently, collecting and interpreting that information at commercial scale has been difficult. At Nectar, we’re working to make it more practical. In this article, we explain some of the ways we do so—and why you should take the results seriously.

Monitoring mites at scale

Alcohol washes are an established, valuable method for monitoring varroa, giving beekeepers a practical basis for making management decisions. Extension guidance often recommends recording results and sampling before and after treatment to evaluate effectiveness.

But doing this consistently across a commercial operation is a challenge. Each wash takes time and sacrifices the sampled bees. Further, many beekeepers are skeptical of the value of mite washes given their inconsistent results and inability to capture mites living under capped brood. Most beekeepers opt to not use this monitoring tool at scale.

After two years of research and development in our CTO Nico Coallier’s hives at Miellerie Flavo, in fall 2025 we introduced an alpha version of Meli, Nectar’s audio-based hive diagnostics tool. Meli uses a 15 second audio clip, taken with a mobile phone placed in the hive entrance (with an option to connect it to a Nectar manufactured microphone that is placed in the hive entrance) to estimate mite levels, with the goal of making frequent monitoring practical across more colonies.

That immediately raised a fair question: why should beekeepers trust the result?

Meli’s convenience in allowing for faster monitoring alone isn’t enough—we need evidence that its readings are meaningful.

Building confidence in Meli

Our first approach was straightforward: ask beekeepers to conduct alcohol washes alongside Meli scans, then compare the results.

We started here because we believed that building a dataset of labeled audio clips with mite levels would be the best way to demonstrate that Meli was increasingly reflecting the results of the most common tool for sampling mites. 

But this proved limited – primarily because alcohol washes do not always provide an accurate picture of a colony’s mite population. So, even if we matched those results, Meli would not be accurately capturing a colony’s mite population. 

Instead, we began searching for another solution. Over the past year, we worked with a committed group of beekeepers who conducted thousands of Meli scans, performed alcohol washes, and recorded their treatments. As that monitoring dataset grew, so did the treatment application dataset. This gave us the opportunity to explore another useful way to evaluate Meli: do its readings change in ways that are consistent with what we know about varroa biology and management?

We call this biological coherence.

As we began to compile the data across Meli scans and varroa records, we began to view this as a superior way to evaluate Meli’s accuracy.

What biological coherence means

Biological coherence means that a tool’s results follow patterns we would reasonably expect from the underlying biology.

For example:

• After a high dosage of an aggressive treatment, we would expect Meli readings to decrease. If treatment reduces the mite population, Meli should detect that decline.
• Without treatment, we would expect readings to rise over time when colonies continue raising brood. Mites reproduce inside capped brood, allowing their population to grow.

When Meli consistently shows these expected patterns, it gives us greater confidence that its readings reflect changes in mite pressure. 

A useful monitoring tool should reflect these dynamics and help reveal additional patterns

However, biological coherence is one layer of evidence. It does not, by itself, prove the accuracy of an individual reading or establish that a tool measures the entire mite population inside a colony.

With more than 20,000 Meli scans now collected, we can examine these patterns across a much larger body of field observations.

What we’re seeing at an operation that uses Meli extensively

This chart brings monitoring results and treatment records together. The upper bars show average reported mite levels, with sample or scan counts in brackets. The lower portion shows recorded treatments by type and date.

January through June includes both alcohol washes and Meli scans. From July onward, the monitoring results come entirely from Meli. That distinction matters: a change in measurement method can affect comparisons over time.

Within the Meli-only period, we see an encouraging pattern. After high July readings, the operation treated heavily. Then, we observe mite numbers decline through August and September. As we get into the third month after the July treatments, brood populations stay high (expected in this particular geography), and mite populations grow in October. That sequence is consistent with what we would expect. We would expect that mite levels would decrease immediately after a round summer treatment. We would expect that, after a few months and continued brooding in the hive, mite numbers would increase again. This is evidence of biological coherence.

The chart demonstrates the value of viewing monitoring and management records together. It makes it easier to identify when readings changed, what happened beforehand, and where further investigation is worthwhile.

Comparing summer management with fall results

The second chart asks a more specific question: how do fall readings differ between colonies that received different amounts of varroa treatments?

It groups colonies by the number of OA-vapor applications recorded and then compares their average Meli readings during that same period. Biological coherence dictates  hives that receive more treatments should read lower mite levels and that this should be reflected in Meli readings

Among the groups with one through four recorded applications, that’s what we see: more applications were associated with lower fall readings.

This is another encouraging observation. Meli is showing a relationship between quantity of treatment and mite levels.

The practical value is that beekeepers can begin examining these questions within their own operation. Rather than relying entirely on recollection, they can identify patterns worth checking against starting conditions, treatment timing, and subsequent results.

Meli captures bigger mite drops from treatments known to drive bigger mite drops

A third analysis looked at how Meli readings changed before and after two different categories of Varroa treatments: aggressive treatments, which we generally expect to produce larger reductions in mite levels, and gentle treatments, which we expect to have a more modest impact.

Across roughly 400,000 hives, we analyzed more than 20,000 Meli scans taken before and after treatment.

If Meli is biologically coherent, we should expect it to detect larger reductions in mite levels following aggressive treatments than following gentle treatments.

And that's exactly what we observed.

For context, here's how we grouped the treatments:

Aggressive treatments:

• Formic flash
• Formic Pro
• Apivar
• Amitraz
• Oxalic vapor

Gentle treatments:

• Oxalic strips (Varroxsan and variations)
• Thymol
• HopGuard
• Taktik

What did the results show?

• Following aggressive treatments, Meli detected an average decrease of approximately 0.7 mites.
• Following gentle treatments, Meli detected a slight average increase of approximately 0.1 mites.

The difference between the two treatment categories was 0.83 mites.

In other words, Meli consistently detected larger reductions in mite levels following aggressive treatments than following gentle treatments. The difference was large enough, relative to the variability in the data, that it was unlikely to be explained by random variation alone.

This is an encouraging result because it suggests that Meli isn't simply producing fluctuating readings. Its measurements are responding to different treatment strategies in ways we would expect based on their impact on mite populations.

It's another example of biological coherence: Meli's measurements are changing in ways that are consistent with our understanding of Varroa biology and management.

How we turn increased monitoring into better outcomes

Monitoring more colonies gives beekeepers a stronger foundation for improving their varroa protocols. But collecting the data is only part of the work. To make better decisions, they need to see that information at the right time, alongside treatment history and colony outcomes.

As we gather more evidence that Meli provides a meaningful signal of mite pressure, we can focus more attention on helping beekeepers interpret and use the information they collect.

Historically, both steps have been difficult. Monitoring at commercial scale takes considerable effort. Even beekeepers who consistently collect alcohol-wash data can struggle to turn individual counts into a clear picture of their operation. Knowing the result for a particular hive is useful; understanding how mite levels are changing across thousands of colonies requires bringing those results together.

Nectar is helping beekeepers answer the following questions:

• What is the average mite level across my operation, and which yards are above it?
• How do levels compare with the same time last year—or over the past five years, where records are available?
• How quickly are mite levels rising or falling?
• Which treatment regimens—including the products, number of applications, and spacing between them—are associated with lower mite levels and better colony survival?
• What does my treatment history suggest about the best time to intervene?

The charts below are some examples of how monitoring and treatment records can come together to reveal useful patterns. As Nectar rolls out more of these easy-to-use varroa insights, we will make analysis accessible without requiring beekeepers to spend hours organizing records or building spreadsheets.

Like the chart above, this chart provides beekeepers with their average mite levels over time with their treatment type and quantity overlaid. This chart comes from a Nectar user who has been entering mite levels and treatment types for a few years.

With high mites last fall, and then again into the early winter, they believed they started their treatment too late, with most hives not getting varroa treatment until late February / early March. The beekeeper was able to use this chart to validate discussions they’d had internally in which they suspected that they might have started treating too late last year. This year, with the ability to back up their intuition with these visualizations, they are more likely to consider starting treatment a bit earlier. 

As the owner of a large operation with a middle management team, the beekeeper whose operation the chart above depicts isn’t personally involved in every field activity. He was surprised to see just how long Treatment 1 had been used.

While he was pleased with how well his operation had kept mite levels under control this year, he also recognizes the importance of balancing effective mite management with the risk of developing treatment resistance. His operation aims to avoid relying on any single treatment for too long.

The chart revealed something he might not otherwise have noticed. Armed with this information, he can now revisit treatment protocols with his team, with the goal of limiting prolonged use of individual products, improving treatment rotation, and reducing the risk of developing treatment-resistant mite populations across his operation.

This chart helps a beekeeper observe where mites are increasing or decreasing the most. This can be a foundation on which a beekeeper attempts to assess if there are drivers of higher or lower mites in the yards listed. 

This chart helps a beekeeper evaluate the effectiveness of a given product. It shows mite levels across yards treated with that particular product, helping beekeepers decide if they should continue to include that product in their varroa management protocol. 

More analysis will come that will help beekeepers understand what products are driving lower mite levels, what timing they are best applied, and more. 

The goal is to give beekeepers all the information they need to make informed decisions.

Conclusion

For Nectar, the burden of proof for validating Meli’s accuracy is twofold: demonstrate that Meli provides a meaningful signal of mite pressure, and show beekeepers how to use that information to improve their management. That means making the visualizations easy to understand and connecting them to the practical questions beekeepers need to answer. 

We are laser focused on releasing varroa charts—vetted by commercial beekeepers—that make it easy to quickly find trends in their outcomes data that help them to assess their situation and improve their protocols. We have already seen this happen with beekeepers that jumped into collaborating with us closely from the start. Now, we are ready to extend these approaches to the rest of the industry.