John Murphy talks to Will Pierce, of the Florida-based spectrometer company Stellar-Net.
When you sell fibre optic spectro-meters and you are based in Florida, you have to expect to be compared to the larger company down the road that does the same. But StellarNet’s Will Pierce is not concerned; in fact he welcomes comparison because he believes he makes a better instrument for about the same money. But don’t expect Ocean Optics to agree. (In fact Mike Morris, president of Ocean Optics, has said: ‘We make the best spectrometers in the world at any cost, and we have sold more than 75,000 of them – more than all our competitors combined, I believe.’)
Though rivals, they are far from enemies. In fact Pierce did much of Ocean’s early development work until he parted ways with them more than 10 years ago. Pierce had a background in military technology and avionics, so he liked to make things that were tough. He decided to make his ruggedised spectrometer and Ocean started expanding into selling all sorts of other optical components as well.
Both companies have built up a strong business for themselves and StellarNet is poised to start moving out of spectrometers into larger systems for vertical markets. Both companies have remained friends and Pierce even picked up a nice bonus when he sold his share in Ocean during the buyout of the company last year.
StellarNet was founded in 1991 by Will Pierce as a consultancy designing process control systems. Pierce had a wide experience in defence electronics and avionics. He taught himself to create automatic test programmes while working on a Navy anti-submarine warfare system. He was dubbed ‘Hawkeye’ by colleagues after the character from the TV show M*A*S*H, because he was so good at fixing things that went wrong. He moved into the software field and worked for Citibank, integrating the software for its first ATM network. He moved back into defence electronics, briefly working on laser gyros, until he started his first software company, called Hawkeye Graphics, devising graphic systems for the first personal computers. He later moved to Florida and started working with Honeywell on inertial guidance systems.
In 1993 he first met Ocean Optics, which had just started making its first product – a small fibre optic spectroscope – nearby in Florida. He started working for them as a contractor and then joined the company and ended up part-owner. He designed many things for them, including the first portable spectrometer and the software interface known as SpectraScope. He decided that some of the features of its spectroscopes were not as good as he thought they could be.
He says: ‘It worked OK, except you couldn’t heat it up or cool it down and you couldn’t drop it. Those were important things and caused problems. I built something that was ruggedised. In 1995 we parted ways, because they were not interested then in doing something that was extreme.
‘I decided to set up my own company and everything we sell, we make. You can drop them, vibrate them and you can heat them up with a heat gun. As people move along, they come across problems and they come to us. We tell them that we can sell them something for more or less the same price, except you can drop it. Also, the original software was in DOS with a Windows-like interface. I redid it in Windows and that became our SpectraWiz in 1998.
‘I stayed friends with the Ocean Optics people and every year I went to the shareholder meetings. In fact, I was the largest shareholder except for the four founders.’
Unlike Ocean, StellarNet has stuck pretty much to spectrometers. Its flagship model has many configurations of optics, such as a concave grating. Pierce also has an infrared version using a Sensors Unlimited detector, which he says is superior to and less expensive than those available from other suppliers.
One of its big selling points is that all the interfaces use the same driver, using either parallel or USB 2.0 ports. The software is designed so that many different detectors can feed into a single screen interface and display in the same graph. Pierce says that, for many customers, it is the software that draws them to StellarNet, even though he is just giving the software away to sell spectrometers.
He says: ‘SpectraWiz is like a Swiss Army Knife of spectroscopy measurement; it has everything that you can think of inside it, including a neural network. Some people come up to me at shows and say they would like to buy the software. They are often surprised that the unit is about $2,750 and you get the software for free.
‘We are focused on having a core piece of software that does everything and, on the side of it, there are two other programmes that can be customised to suit your needs. If none of those apply, we can set a contract to customise something like LabVIEW, for example, to do exactly what it is you want to do. We give them the source code for the side programmes so people can modify it, which researchers like. We don’t build the overhead of developing special software into the price of the hardware, so we have to charge separately for it.
‘One of our mottos is “the software is the instrument”. Once you get over the spectrum being there, what do you do with it? There are about a dozen parameters that you would want to measure, or there are the chemistry guys who just want to know what the concentration is of various chemicals. You can also use a spectrometer to look at plasmas, and we make the world’s first truly portable laser-induced breakdown spectroscopy system that fits in a briefcase and runs on a battery for a week.
‘We are not in the business of vertical markets at the moment. We just want to deliver enough hardware and software so that someone can take that over and go for a vertical channel. Otherwise we would spend all our time going up vertical channels.’
The low cost of the instrument has meant that StellarNet has to cut costs and the sales strategy is one way that economies are made. He has built up a network of dealers in most parts of the world. Some are larger distributors, like Laser 2000, and others are small, local companies. They are highly trained to support the instrument, which has allowed Pierce to keep the core US staff small and focused on the US market. The low sticker price demands a low-cost sales model. In the US it sends out instruments on a sale-or-return basis. Anyone who wants one can have it shipped and then just pay 10 per cent of the costs as a handling charge if they want to return it. Pierce says that, so far, nobody has. But this does not seem to be working in Europe.
Pierce says: ‘In the UK, for example, they demand to have the distributor come to the site for a demo. Our rivals send out guys to do a demo and we ask them how they can afford to do that when the customer doesn’t know how to drive it and you are supporting him all the time. The distributors tell me that is the way it’s done in Europe, and I’m amazed. I have been trying to train guys in Europe not to take it. I’ve told them to tell the customers to bring their samples in and do it at their offices.
If there is a potential of a large OEM order then we do spend a bit more time hand-holding them, but for the one-off it does not make sense in the low-cost arena.’
About half of StellarNet’s customers are in academic research. The other half are in industry, mostly doing online quality control on colour. The food industry is a fast-growing market. About half this market is photonics – measuring the transmission of light through an object, for example – or spectra-radiometry, which can be anything from testing sun tan lotion to reflectance from vegetation. The other half is chemistry applications, such as material analysis. Pierce says the main attraction to industrial users is the fibre coupling, which allows them to do QC on their production lines.
Pierce says that there are many potential customers out there who have not even thought of using fibre-coupled spectrometers yet. He says: ‘There are OEMs out there who are highly knowledgeable about their fields and are looking for components to make their products better. We get two or three of those every day.’
StellarNet was recently working with scientists from the University of Georgia, who had created a spin-off company that was trying to measure, non-destructively, the amount of soluble solids in an onion by shining a white light through it.
He says: ‘They sorted out the onions with a higher percentage and sent the rest to market. The next year they replanted their fields with the sorted onions and each year they will get millions of dollars more for their onions because they are paid by weight. The instrument sold for $35k.’
StellarNet has produced a wide range of spectrometers and has started adding other accessories, such as light sources. Pierce found that it was difficult to find a light source that was reliable and long lasting and he had to spend a lot of time supporting instruments where the light source was at fault. So he designed his own and started supplying a full turnkey instrument. He has also made reinforced cables that meet the same rugged standard as the instruments. Another accessory is a multiplexer for the InGaAs detectors. There are very expensive, so he designed a system that can look at four different locations using a single detector.
Pierce says he does not need to know what customers are going to do with the device if they prefer not to tell. He needs to know what type of measurement they want to make and what kind of sample, and from that he can configure the component to do what they want. He finds that if something goes wrong it is usually ‘cockpit error’.
StellarNet is a lean company, with only about a dozen people at its main office. Pierce likes to keep the core staff small, because it helps keep communications simple. He tries to make sure that once an instrument is shipped, it stays shipped and believes that if you get it right to begin with you don’t get the support calls. He finds that it takes about nine months to train new people. He also wants to hang on to his dealer network, because he has found it takes several years to train them. His staff don’t just do sales; everything is made in-house and, like the crew of a submarine, all the staff can do almost any job. His hobby is undersea exploration – which may be where he got the idea from. He has tried getting boards built by another company, but found that so many boards failed and had to be rectified that it was easier and more reliable to build in-house.
Turnover has reached about $3m this year and has consistently grown. His target is to hit $10m in the next few years and he has started increasing the marketing activity by attending more trade shows. He often finds he sells as many instruments to other exhibitors at trade shows as he does to delegates.
He also wants to increase turnover by moving up the food chain. He says: ‘We are coming up with a bigger solution that you can charge more money for. We are going to build some application-specific configurations that are an all-in-one solution. We are looking at systems for food, analytical systems and so on. We will need to go into the journals for those specific markets. We find that we can sell something for $25k, but if we add into it some specific database information so that it’s useful in a specific vertical market, then we can sell the same piece of equipment for $75k. It’s only software and databases. It is competing with other systems that do the same thing but cost many times more. Customers are ony interested in results.
‘This has been in the game plan all along; it just requires the collection of possible product types to build because there are so many.’