It can take a decade of hard work to bring a new orchid to market.

While the rewards can be significant - the global orchid market is worth hundreds of millions of dollars - the competition to produce the next gorgeous flower is intense.

Which is why, in the race to develop new orchid types, the laboratory is at least as important as the greenhouse.

Centuries of human intervention - selective breeding and propagation - have made the genetic background of many commercial orchids a "disaster", according to leading Dutch orchid breeding firm Floricultura.

That means it is extremely difficult to predict what characteristics a new plant breed might have.

But by developing genetic markers for particular traits - colour, shape, disease resistance, flowering longevity and so on - Floricultura and its competitors can try to speed up the process of selective breeding.

Instead of waiting for a newly bred plant to flower in three years' time, the breeders can apply genetic screening techniques to very young plants and discard those that don't meet their requirements, right at the start of the process.

"If a few thousand cross-breeds [come] from the lab, we can screen them based on the marker and just select the ones that have the marker that you search for," says Wart van Zonneveld, Floricultura's research and development manager.

"It's an indication of a certain trait that you want, or you do not want, depending on what's easier to find."

So-called "novel breeding techniques" are closely guarded secrets. Each company develops its own genetic markers and processes because that's what allows it to develop unique varieties.

"We keep it to ourselves because it's lots of investment," van Zonneveld says.

"It's still breeding, you have to make a cross, and we cannot just pick out a piece of DNA and put it back that easily," says Paul Arens, ornamental plant breeding researcher at the Netherlands' Wageningen University.

He and his colleagues have carried out research for a Dutch government-backed initiative that shares information with participating companies.

"The foundation is still what we have been doing for 100 years already. You take two plants, you look at their characteristics, and you make a cross. But [the breeders] have white lab coats, [and] they're doing all kinds of research with markers, with genomics, on plant health."

Genetics are also used to protect the intellectual property in the new variety itself – in Europe through breeders' rights, and in the United States through patents.

"If a company makes a new orchid, then [it] would like the sole right to commercialise this orchid," Arens says.

"Otherwise, somebody else can just buy it in the shop, multiply it and sell it himself.

"But the breeders' rights researcher has to make sure that a new variety is distinct from anything that's already in the market... it has to be distinct, it has to be stable, and it has to be uniform."

Breeders' rights and patents are granted based on physical descriptions, not DNA analysis, but it's essential to compare the new plants with similar products to determine whether they qualify for protection.

DNA analysis is a powerful tool in determining which plants the new variety should be compared with.

"It's just like what we do in forensic science. You run markers that are at different positions in the DNA and that give you a pattern, and then you have a chance to match it or not," Arens says.

Floricultura don't sell to the public, or even to garden centres. Their business is to develop and produce new varieties, which they sell to cultivators who grow the plants at scale.

They have more than 180 varieties in their catalogue, but several hundred more in development, because the demand for novelty and development never ceases.

"You can't stop, because it takes so long to develop new varieties," says Stefan Kuiper, the company's breeding manager.

"You have to go on, [or] you will be behind the rest."

After genetic screening and initial selection, the plants (the first attempts at a new variety, siblings from the parent orchids) take around three years to grow, first in lab conditions and then in greenhouses, but there are still years to run in the development stage.

Breeding, says Paul Arens of Wageningen University & Research, "is the art of throwing away", discarding those plants which don't match your ambitions, but it is also the art of multiplying what's left.

Because the next batch of plants won't be siblings, they will instead be exact copies of the ones which survived the selection round - clones.

"In the beginning, everybody had the seedlings, so the crossing and then the seed pods give plants, but we at Floricultura introduced meristems," Stefan Kuiper tells me.

Meristems are the cells that allow a plant to continue growing throughout its life, and it's these that are used to clone surviving plants.

Stefan can't explain more about the technique they use – like the genetic research, it's a trade secret.

However, the cloned seedlings are cultivated and grown, again over the years, to another selection point.

The cultivation of orchids is a resource-heavy business. The plants need reliable heat, light, water and nutrients, over many months.

The application of genetics and other techniques can only speed it up so far. Ultimately, you have to let the plant grow, confirm the characteristics – flower shape and size, colour, the number of stems, resistance to disease and so on – and then make another selection.

That process transports the young plants by airfreight to India and by lorry to Poland, before they return to the Floricultura site in Heemskerk in North Holland, where there are more than 7 hectares of greenhouse space for both development and production.

Rainwater is harvested from the greenhouse roofs, and in response to changing weather patterns, the company is now beginning to recycle that water and the nutrients it contains for secondary use.

Wart van Zonneveld proudly showed me their geothermal well, which pumps water up from 3km below the surface at 102 °C.

It provides so much energy that they are exploring sharing it with the local council for district heating projects.

It's not only the monitoring that is automated. In the vast greenhouses, trays of plants shuttle around on rollers, which deliver them to the next stage of cultivation in sequence.

There remains one task which, at Floricultura at least, is reserved for humans.

Whilst the tools for developing new varieties, cloning new plants and assessing the results have all been transformed by technological innovation, the decision on which varieties, after nine years of work, make it into the catalogue is still made by Stefan Kuiper and his colleagues in person.

A plant can have tick all the genetic boxes and produce all the right traits, but it has to be beautiful to sell – and that's a judgement made by people.

"Breeding is a little bit [like] gambling", Kuiper says, and for now, that human element remains.