Skip to main content
Premium Trial:

Request an Annual Quote

Cover Story Affy Gets Affable


With competitors coming from all directions, the microarray queen cozies up to customers

By Aaron J. Sender & Marian Moser Jones


By the time its GeneChips began flooding the commercial market in the late ’90s, Affymetrix had already developed a reputation for Microsoft-esque arrogance. “We didn’t have a customer-friendly reputation. And we weren’t a customer-friendly organization,” says CEO Steve Fodor. “We didn’t know how.”

The bad rep was Fodor’s penalty for being a visionary with tunnel vision. He had merged a semiconductor manufacturing process with complex biochemistry to create a gene expression microarray market out of thin air. And he had built a multimillion-dollar company, and positioned it to take the biggest chunk of what is predicted to become a multibillion-dollar market. But in his rush to dominate the world of nucleic acid analysis, Fodor and his company had overlooked one crucial element of success: the customer.

Affy’s first customers, the early-adopter variety who had invested valuable resources on a still unproven technology, expected to have a voice in shaping the novel tool to their needs. They had concerns. They were troubled by the reproducibility of array results, the effectiveness of the algorithm used to turn readouts into meaningful data, the inflexibility of the platform, and the lack of access to the sequence of the oligonucleotide probes on the chips.

But their complaints fell on deaf ears. Affy’s customer service was iffy at best. And its genial-sounding nickname, Affy, began to be used more with irony than affection.

“When you’re the only game in town, there isn’t that competitive pressure to listen to your customers and do what it is they say they need,” says David Pritchard, who, as senior VP of Neomorphic last year, negotiated Affy’s $70 million acquisition of the bioinformatics company.

In recent years, however, insiders and outsiders came to realize that things would have to change. Dozens of companies eager to chip away at Affy’s market share began tuning their technology to address the issues Affy was neglecting. Protogene, Agilent, and Motorola, for example, began touting customizable chips. Incyte rolled out chips with unique content as well as paid access to its probe sequences. Nanogen promised faster hybridization directed by electrodes. And unlike Affymetrix’s GeneChip, a closed system that requires customers to buy an Affymetrix scanner and software, competitors began offering open-platform chips that could be read by any standard scanner.

Pressure was especially intense because it wasn’t just coming from startups but from established companies, such as Motorola and Corning, whose yearly revenues far exceed Affy’s total market cap.

“There’s a certain swagger, a certain arrogance that market leaders take on. And Affy can’t afford that luxury,” says Pritchard. By 1998 it was all too clear that change was in order.

Finally, the catalyst came in the form of Sue Siegel. Fodor’s hope was that her outgoing, friendly personality and experience at a customer-focused firm would help give his compay the image makeover it so desperately needed.

Affy’s Attitude Adjustment

Sitting in an oversized conference room chair at Affymetrix’s Santa Clara headquarters, Siegel, 41, beams from behind her rectangular black-rimmed glasses. Shy of five feet, Siegel is a challenge seeker and an athlete. Confessing to being an avid volleyball player, she laughs, “I could probably walk right under the net.” What she lacks in stature, though, she makes up for in charisma — Siegel’s personality has a way of filling a room.

Siegel joined Affy as senior VP of marketing and sales from Amersham Pharmacia Biotech in 1998 and quickly jumped to take the reins as president. Her mission was to regain customer loyalty and trust. And she had to start from the bottom. “We were not the most popular company,” she says.

The first order of business was to set up a sales and marketing infrastructure. Siegel went on a hiring binge. The company had grown to more than 200 employees but “when I joined, I can count on my 10 fingers how many people existed within what we call the ‘commercial organization.’” Within six months there were 60. She built a customer service team, a sales and marketing force, and shipping and receiving departments.

In the past, customers often complained that they weren’t getting chips fast enough and that they could not count on reproducibility from one batch to another. “Customers were saying sometimes they were waiting four months to get chips,” says Siegel. Affy needed to get a full-scale manufacturing process up and running with improved quality control. The original Santa Clara, Calif., production site was too small and it lay on a fault line.

In January 1999, on 10 acres of solid Affy-owned ground 120 miles northeast of the corporate offices, a new three-acre plant began producing chips. (Serial-biotech-company-founder and the brain behind Affymetrix’s parent company, Affymax, Alejandro Zaffaroni had the foresight to suggest that the company buy more land than necessary to allow for expansion.) The company hired Steve Karas, a pharmaceuticals manufacturing executive with 23 years at Hoechst, to run the plant and streamline the chip-making process. “Now,” says Siegel, “from the time we receive the order until the customer gets it, it’s approximately four days.”

Siegel also began a habit of interviewing her technical support employees. “The technical service folks are the first ones to hear issues with customers,” she explains. The most valuable information often comes from new employees, not yet immune to the products’ shortcomings, Siegel says. “I followed one person in particular and I’ve asked him over and over again, ‘OK, so tell me the worst about our technology.’”

To further foster what Siegel calls “customer intimacy,” she set up focus groups and user meetings. Siegel is a devotee of marketing guru Geoffrey Moore’s Crossing the Chasm model of technology


adoption. The idea is that the techies who adopt a technology early on are willing to accept a product that doesn’t yet do exactly what is promised. Mainstream customers, on the other hand, want a product with all the kinks already worked out. Between these two groups is a large gap, or chasm. Crossing it, the theory goes, requires paring down the broad vision into a single, complete product targeted at a specific market segment, and establishinga “beachhead” in that segment. For Affymetrix that beachhead lies in pharma and biotech’s early-stage drug discovery efforts.

With its beachhead conquered, a company is in what Moore calls the bowling alley. (The theory is not for those with a distaste for mixed metaphors). From here, the aim is to use existing territory, or pins, to knock down the next one. For Affymetrix that means moving down the drug development pipeline through toxicology, pharmacogenomics, clinical trials, and finally diagnostics. “If you can get your current customers happy, and they’re the ones that have been using the technology, they’re going to know what your next customers are going to need,” says Siegel.

Following this model, Affymetrix has started in recent months to behave like a prodigal spouse toward its customers, listening to their feedback at user meetings, and pledging to make changes in order to keep them from straying.

For instance, when Affy discovered in March that its U74 murine chips contained incorrect sequences, it responded immediately, flying out representatives to meet larger customers, and then announcing it would replace the defective chips completely within six weeks. Although customers weren’t happy to hear about the error, many were pleasantly surprised by the vendor’s candor and its speedy response.

In July, Affy complied with requests for flexible content with CustomExpress, which allows users to design their own arrays by choosing up to 16,000 probes from the prefab GeneChips. CustomExpress program customers must order at least 90 chips, and pay a design fee for smaller orders, so it is primarily a boon for larger customers in pharma and biotech. But another new customer-friendly offering, the web portal NetAffx, allows users to search public and private gene databases for information about genes represented on the chips. The portal is powered by a search engine based on Lion Bioscience’s SRS, and can be accessed by any existing or even potential customer.

Finally, in August, Affy reacted to a long-standing concern about statistical flaws in its GeneChip algorithms by unveiling a new method — one that has flexible parameters and no longer assumes that gene expression data fits a normal distribution. In accordance with Affy’s new Glasnost approach, its scientists plan to publish the algorithm in a major scientific journal by the end of the year.

Now, the market is beginning to take notice of Affy’s new attitude. “Customer support has been amazing with Affymetrix recently,” says Gregory Khitrov, director of the Rockefeller University core microarray facility, who has been an unabashed critic in the past. “It’s a lot better than it used to be. It’s night and day.”

Yin and Yang

Siegel, whose customer-minded yin complements Fodor’s visionary scientific yang, insists that this is just the beginning of Affy’s era of openness.

But she is still second in command. Next to Siegel, Fodor strikes a casual pose at the conference table in blue jeans, a dark button-down shirt, and wire-rimmed glasses. He tends to talk about the grand implications of his company’s technology and how his venture has grown from a handful of scientists with a dream to a 900-employee company that has changed the way researchers view DNA analysis. It’s clear that he still dictates the technology’s development.

For instance, many researchers, loath to trust anyone else’s calculations, have long begged for access to Affy’s probe sequences and algorithms that convert raw data into gene-expression levels. But Fodor won’t budge.

Siegel insists that even these last vestiges of secrecy will disappear with time. “Our intent is to be more and more open. And you’ll see much more of that coming up,” she says. “The more we can open this up in terms of the information, and the more academics can understand everything behind it, the better off it ends up being for everybody. So you’ll see us continue to push in that direction.”

When asked about whether the company would actually go so far as to allow people to improve upon Affy’s algorithms, Linux-style, Siegel doesn’t answer directly. But she gives this much: “We don’t think we’re the only ones that can get it right.”

“In fact we know we’re not,” Fodor adds. On this point at least he seems to have been tempered by Siegel’s insistence that customers deserve a seat at the table.

Corporate CULTure

What has driven Siegel to spend her days, as well as many late nights, instilling Affy with this new philosophy of openness?

After just a few moments with the pair, it becomes clear that Siegel, like so many others, has been won over by Fodor’s infectious zeal for transforming drug discovery, healthcare, and even marketers’ understanding for how genetic profiles affect people’s preference for things such as beef, chocolate, and perfume.

When Fodor first began courting Siegel to join Affymetrix, “I was not saying yes, let’s put it that way,” she says. “I was in a very nice company. They had taken very good care of me. And here’s Affy, a small company, high risk, a new area.”

Siegel had made up her mind to stay at Amersham when she and Fodor shared a flight from Boston to San Francisco. “He gave up his upgrade to sit with me in coach in a middle seat way at the back of the plane. We had six hours of the most intense discussion of Steve’s vision about what the technology could be used for, absolutely nonstop. When I landed, I can tell you, I had the biggest headache,” she says.

Fodor had been trying to seduce Siegel for three and a half months. “What really pushed me over was that six-hour plane ride,” she says. The way Siegel describes her conversion sounds more like a religious awakening than a corporate defection. “Steve was talking about not just healthcare, but the opportunity to use the technology for the good of humankind.”

At Affymetrix headquarters this almost cultish faith in the far-reaching benefit of the technology is pervasive. Huge green and red tiles suggestive of microarray images plaster the walls as if to declare the GeneChip larger than life. On a clean white coffee table in the lobby, copies of the annual report fan out to greet visitors. A saucer-sized blue retina peers out from the report’s cover, emphasizing the cryptic slogan “Every Living Thing.” Minutiae such as revenues or the business plan are confined to two pages at the report’s end. Readers flip first through a Technicolor photo album featuring a human hand, chromosomes, DNA, the multi-ethnic faces of the human family, and, of course, GeneChips.

This is a company where piety trumps pecuniary matters to such an extent that employees repeat mantras like “We’re excited about the opportunity to alleviate human suffering” and refer to their work as “a moral obligation” with absolute sincerity. The uninitiated visitor needs to suppress groans.

Too little, too late?

Affymetrix is counting on its new and improved relationships with its customers to accelerate a new world order. With an installed customer base, the chipmaker has something that its competitors don’t: access to feedback from hundreds of users who have already applied its technology to real science. “Our customer base is extremely valuable,” says Siegel.

But are Affy’s affections too little too late? Instead of rallying around the market leader to protect their investment, some neglected customers have begun rooting for emerging competitors.

“There are a lot of different possibilities out there and I don’t think anybody at this point can definitively say which of these technologies is better than the other,” says George Busby, administrative director at the Harvard University Center for Genomics Research. His center uses both Affymetrix and Incyte arrays, but that may change. “It’s far from a settled issue,” Busby says.

To stay on top, Affy will have to win in head-to-head comparisons. “We’ll put them side to side and analyze the pros and cons of each technology,” says Busby. “We’ll compare signal intensity, background sensitivity, dynamic range, quantities of material required, and the possibility of the use of RNA amplifications.”

Siegel relishes the challenge. “Data is everything,” she says. “What this will do is force the issue of ‘Let’s get to the data.’”

One competitor that has gotten Affymetrix’s attention is Motorola. In August, Motorola released a chip with 10,000 human genes and promised more chips to follow. Despite Affy’s hold on scientific community mind share, Motorola insists it’s much too early to crown a microarray queen. “This is not Intel in the microprocessor world, not quite yet,” says Motorola corporate VP and life sciences manager George Turner. “They have a large installed base,” he says, “but only in what we call the SAM of the TAM.”

The TAM is the total available market, the SAM the part of the market currently being served. According to Turner’s calculations, 80 percent of the market is still wide open. In other words, Affy has the dominant share of the 20 percent that is currently buying microarrays. “There’s still a very large self-spotting market. We believe that those folks have either not found a cost-effective or a compelling enough solution,” says Turner.

Harvard’s Busby agrees. “There is certainly the possibility for other players to come on to the scene and, if they have a really good idea and if they’re able to market successfully, take over a large share of the market.”


The communications giant even has a facility next door to Affy’s headquarters, leading one Affy employee to joke about spies with binoculars. Actually, the facility is unrelated to Motorola’s life sciences arm, but still, Fodor admits, “We are always very paranoid.”

Motorola and other competitors seek to cut into the market by homing in on two weak points: Affy’s reliance on the expensive and inflexible photolithographic process, and its insistence on using public data as content for its chips. Motorola, which has licensed Incyte’s content, creates oligos first, checks the sequence, and only then deposits them on an array. Affy, on the other hand, builds the oligos right on the chip, which makes it difficult to verify the probe sequences. “Look what happened with their mouse arrays,” says Motorola VP of commercial operations Mike James.

Also, photolithography requires that chip design be etched in silica. Each design change requires a complete new set of masks to filter the light needed to pattern the oligos on the array. Affymetrix would not say how much a set costs, but if similar masks in the semiconductor industry are any indication, it could be as much as a hundred thousand dollars.

Agilent, which inherited parent Hewlett-Packard’s proprietary inkjet technology, hopes to capitalize on this perceived weakness. The Palo Alto, Calif., company has also put a 10,000-gene human catalog cDNA array on the market, which it prints using its inkjet technology, and plans more by the end of the year. It also runs a growing custom-array business with oligo-based printed chips. And because the content deposited on the chip is determined by a software program, not a physical mask, “we can change the design almost instantaneously,” says Agilent marketing manager for bioresearch Wilson Woo. Agilent also touts proprietary content it has licensed from Incyte as an advantage over Affy’s chips. (Affy, meanwhile, has spun out Perlegen, which will scan 50 individual genomes, chromosome by chromosome, to generate its own content.)

Corning, another biggie with eyes on the market, has failed to meet several announced release dates for its high-density arrays and refuses to comment about its progress. The Harvard center, which has beta-tested the arrays, says the fiber optics company has some issues still to work out.

There are also about a hundred biochip or array startups to keep track of. Companies such as Protogene, which stresses flexibility, PamGene, which makes porous aluminum oxide arrays combined with flow-through incubation for speedier experiments, and bead-based companies such as Illumina and Lynx are just a handful of the new entrants that hope to develop something better than the GeneChip.


Lynx CEO Norrie Russell even prophesizes the end of the microarray era. “Getting some general qualitative picture about how a gene is upregulated as a result of disease, that’s not enough now,” he says. “What we need to move on to now is quantitative gene expression analysis.” Whether ultimately Lynx’s microbead technology or some other innovative platform will be the answer is yet to be determined. But so is Affy’s position in the market.

“We need to worry,” says Siegel, who is well aware that, as IBM learned the hard way in the PC industry, the smallest competitors can often be the biggest threat. It’s a matter of who makes the right strategic decision at the right time.

“With regard to other companies, we’re going to be alert, be it a small company or a big company,” she says. “It doesn’t matter who it is or what the technology is — whether it is in the microarray space or any other technology that potentially could substitute it — we’re going to pay lots and lots of attention.”

After all, in a market that could expand to 10 times what it is today, who knows where the chips may fall?

The Scan

Suicidal Ideation-Linked Loci Identified Using Million Veteran Program Data

Researchers in PLOS Genetics identify risk variants within and across ancestry groups with a genome-wide association study involving veterans with or without a history of suicidal ideation.

Algorithm Teases Out Genetic Ancestry in Individuals at Biobank Scale

Researchers develop an algorithm known as Rye to tease apart ancestry fractions in admixed individuals at a biobank-scale, applying it to 488,221 UK Biobank participants in Nucleic Acids Research.

Multi-Ancestry Analysis Highlights Comparable Common Variants at Complex Trait-Linked Loci

Researchers in Nature Genetics examine common variants implicated in more than three dozen conditions, estimating genetic effect similarities across ancestry tracts in admixed individuals.

Sick Newborns Selected for WGS With Automated Pipeline

Researchers successfully prioritized infants with potential Mendelian conditions for whole-genome sequencing or rapid whole-genome sequencing, as they report in Genome Medicine.