In late February, only one day after he’d signed a deal involving the sale of most of the shares of online genealogy platform MyHeritage to the private equity firm Francisco Partners for $650 million, Gilad Japhet was already immersed in another project.
He was getting ready to launch a new tool that would animate the faces from old photos. He realized that such a technology would be met with mixed reactions, and thought that the launch should take place in a friendly environment, allowing it to go viral. The best place to launch such a feature, he believed, was the biggest annual genealogy conference in the world, called RootsTech.
MyHeritage took on a major sponsoring role in the conference in exchange for tens of thousands of dollars, and was invited to deliver part of the keynote speech. Japhet’s people presented the new tool to 250,000 attendees over Zoom. “I called this tool ‘Deep Nostalgia’,” says Japhet. “We emphasized to genealogists that only photos of deceased people should be used, not of living ones, so that the product would not be misused. Many users have photos of their parents or grandparents who died before video was invented. Seeing them move their faces gently, with a dreamy and reflective look, is a moving experience.”
In Egypt there was a fatwa – a Muslim religious edict – determining that it is allowed to animate people using the MyHeritage app, even though it was developed in Israel, because the animation arouses respectful sentiments
The launch unfolded exactly the way Japhet had hoped. Attendees were excited about the product and shared the animations on social media, creating massive traffic for the company’s website. Massive media coverage followed and MyHeritage’s servers were soon overwhelmed. Just before midnight, 30 of the company’s employees gathered in its offices, along with Japhet, starting five days of round-the-clock efforts to bolster the website’s infrastructure.
“The margin between success and failure is razor-thin,” explains Japhet. “All of a sudden, the website had millions of visitors, trying out everything. It’s like opening one successful store in a mall and getting a wave of buyers all at once, asking, OK, what else do you have to offer? We had one feature that allows you to record a photo’s story. It wasn’t in much use, with only about 200 uses a day. But during the launch of ‘Deep Nostalgia’, one developer told me there were 40,000 recordings made in a single day; millions of users began constructing basic family trees on our website, compared to 50,000 on regular days. If the website collapses during such a period, you lose a lot of money, with people calling it a waste of time and posting negative feedback.”
The figures speak for themselves. In the first five weeks since the launch, MyHeritage has animated 72 million photos, and the company’s iPhone app has climbed to the very top of the app store. In Egypt there was a fatwa – a Muslim religious edict – determining that it is allowed to animate people using the MyHeritage app, even though it was developed by Israeli companies (MyHeritage based its tool on technology developed by the Israeli start-up D-ID), because the animation arouses respectful sentiments.
The launch was also translated into the company’s coffers: bookings from subscriptions in March rose by 83%, compared to the same month last year, with 100,000 new subscribers in the first quarter, twice as many as in average quarterly periods. Japhet believes that even though a drop in the tool’s viral character can be expected, the current rate of several hundreds of thousands of animations per day should continue for the rest of the year.
- 'Spooky' AI Tool Brings Dead Relatives' Photos to Life
- Artificial Intelligence Helps Identify Authors of Dead Sea Scrolls
- Israel’s MyHeritage Defies the Slump in the Home DNA Testing Market – for Now
With such an impact on its business, it’s hard not to sense that something was missed in the deal with Francisco Partners. The deal was estimated to be worth $600 million, but included several milestones that raised this to $650 million.
“We surpassed the goals of this deal by the end of the first week, the buyers made a killing, but I and my employees still have shares in the company, and we’ll take it to another exit which will be larger and happen faster,” he says. Private venture funds often acquire companies in order to raise their value and then sell them or take them public. Employees now hold 27% of the company, with 20 of them becoming millionaires last week with the signing of the deal and the receipt of regulatory approvals.
Japhet says he wasn’t eager to make an exit, but wanted one anyway in order to reward his veteran employees, who see a plethora of sales and public offerings by Israeli start-up companies. “I provide my employees with financial and family security and proof that the company they work for is successful,” he explains. He sees the company as a life project and doesn’t want to risk it now by going public. He doesn’t ever want to leave it either.
Photo enhancing tools are helping MyHeritage attract a younger population to their platform, they also distinguish the company from its greatest competitors, Ancestry.com and 23andMe
All about my mother
The story of MyHeritage began with a staple of Israeli culture – a genealogy school project. Japhet was in seventh grade when he was asked to conduct a short interview with his mother. Sara Japhet, a laureate of the Israel Prize for Biblical Research and the first woman in Israel to obtain a Ph.D. in Bible, was the director of Israel’s National Library and the director of the Institute for Jewish Studies at the Hebrew University. To this day, aged 86, she’s an active lecturer who publishes two books a year. Japhet was fascinated by his mother’s childhood stories about the orange groves of Petah Tikva and about her doings on the day Israel was established in 1948. He submitted to his teacher a 40-page interview about his mother’s history, defining the experience as “unfinished business.” It took him 17 years to come back to it.
Japhet built the first version of MyHeritage for his own use. In 2000, at the age of 30 and after 8 years as a high-tech engineer, he wanted a break, deciding to devote half a year to passion projects.
“My wife was quite alarmed when she realized that I wasn’t on vacation but was obsessively investigating family roots, day and night,” he remembers. Japhet traveled to interview relatives, scanned old photos, but the things he wanted to do with his family tree did not exist at that time. He decided to use his skills as a software engineer and develop the tools he needed himself. “It annoyed me, for example, that when you looked at an old photo you were told that your grandmother was the fourth person from the right in the third row from the top. I wanted to tag her. Today it sounds trivial, but this was the first software in the world which used tagging.”
Japhet wanted to introduce facial recognition into his software. His grandfather, Haim Japhet, was personal assistant to Zionist leader Henrietta Szold and among the founders of social work in Israel. The family had no photos from his youth and his grandson wanted to obtain some. He was convinced that they existed in photo albums belonging to other people, and believed he could locate them if everyone had the ability and motivation to upload their old family photos to a shared website, tagging the people they could identify. The system would be taught to identify faces, so that it could look for relatives of users in photos belonging to others. In 2003, financing it himself, he set up a company called Inbal Genealogy Technologies, later called Inbaltech (after his eldest daughter) and then MyHeritage.
Old photos, he claims, are better for facial recognition algorithms. Taking photos used to be a festive event, with people facing the camera frontally, looking serious. However, since facial recognition wasn’t his field of expertise, Japhet located a start-up in German whose technology was being used for detecting terrorists at airports. In 2004, while still an inexperienced entrepreneur, he flew to meet the start-up’s founder, telling him he wanted to use the same technology over the internet, in order to look for genealogical roots.
“I provide my employees with financial and family security and proof that the company they work for is successful"
“He fell off his chair,” says Japhet, who also had an interesting proposition. He told the founder that he’d probably like to know how a terrorist would look in five years, and that MyHeritage could teach his system how to do that. “In genealogy, people tag their relatives from a young age and into their 80s,” he explains. Japhet promised to send anonymous collections of photos with no identifying features, which would enable that start-up founder to develop algorithms for following the aging of a face. In exchange, he got the technology he wanted for free.
From a business perspective, there was no certainty that a genealogy start-up would succeed. But Japhet had a feeling that if this area fascinated him, there must be millions of others like him. But investors didn’t think so, to say the least. In 2005, two years after the start-up was founded, the money he brought from home ran out, when he turned from a successful breadwinner into a debtor, it was time to turn to venture capital funds. Japhet kept a list of more than 30 funds that rejected him, as well as others that dragged things out before deciding not to invest. Ultimately, he got $700,000 from Yuval Rakavy of BRM and from investor Aviv Raiz. “They slaughtered me, leaving me with 50% of the company, at a negligible valuation in today’s terms, but it was the best deal I could get,” he says. Rakavy and Raiz continued to invest in the company and retained some of their holdings.
Japhet still keeps the first business plan he wrote in February 2004, consisting of 100 pages printed in English, with diagrams and color photos, something considered quite fancy in those days. The plan included an in-depth analysis of 39 competitors in Europe and the U.S., with details of their strengths and weaknesses, including some companies that MyHeritage ended up buying.
Japhet had estimated that his sales for 2007, the company’s fourth year, would amount to $1.7 million. The financial part didn’t quite pan out, since MyHeritage gave away its products for free and didn’t start selling products for the first five years. It later adopted a freemium model, in which the basic service is free, while more advanced features, such as creating expanded family trees (with more than 250 people) or more complex searches and access to historical documents require an annual subscription. In the company’s sixth year it had revenues of $3 million, with $8 million the next year. “Usually, such plans are like science fiction. You think you’ll make a certain amount but don’t really believe it. I look at it now and I believe I lived up to the plan,” he says laughingly.
Currently, 920,000 users pay annual subscription fees for MyHeritage’s services. Bookings from subscriptions rose by 28% in 2020, reaching $137 million. The company’s total bookings, including from sales of home DNA testing kits and coronavirus testing services, are estimated at $268 million in 2020. MyHeritage has been profitable since 2012, and its EBITDA from subscriptions last year was $28 million.
Rising above the noise
Photos were always an integral part of MyHeritage. It's great break came in 2006 due to a humorous algorithm that compared users on social media to celebrities that resemble them. The feature was launched after a glitch in the training of the algorithm for face recognition recognized Japhet as the tennis player John McEnroe. The marketing gimmick was called: “Find the Celebrity in You.” It drew in lots of youths, allowing the company to “rise above the noise,” according to Japhet. The gimmick appealed to people’s common superficiality, but it raised interest in the media and afforded MyHeritage an opportunity “to tell what it really does.”
"We’re a company with subscribers and information whose value appreciates with time, unlike the rest of the internet. The reason for our high customer retention rate is that people see their family history as an asset they want to keep for life and pass on to the next generations"
In recent years, MyHeritage has again shown that it knows how to go viral, and dealing with photos has become a growth engine for the company. “Building a family tree and adding names and facts to it is a bit dry. But when you add photos, it becomes more exciting and real. You discover that your great-grandmother looked like you, or you may find a cousin who has the same gestures as you do. You meet yourself in all sorts of versions,” explains Japhet.
In June 2020, at the height of the coronavirus pandemic, the company launched several tools for improving photo quality. These tools are based on technology the company licensed from two foreign start-ups (Remini and DeOldify). These technologies allow you to turn photos that are blurry or pixelated with low resolution into sharply-focused photos, as well as to colorize them automatically.
The results are very impressive if flawed, and were embraced enthusiastically by users. Colorization and photo enhancement are done by artificial intelligence algorithms that infer what faces looked like.
Japhet received the moniker of an Israeli Sherlock Holmes from a French journalist, after taking on as a hobby the location of legal heirs to the owners of art work robbed by the Nazis and held by French museums. He managed to trace eight heirs and restored 20 works of art to them
These days, the company is working on technology meant to repair scratches and holes in old photos. Another algorithm will estimate the date and location in which the photo was taken, based, among other details, on fashion, hair style and furniture, in order to provide extra value for people with unidentified historical photos. Photo enhancing tools are helping MyHeritage attract a younger population to their platform, they also distinguish the company from its greatest competitors, Ancestry.com and 23andMe.
‘The Israeli Sherlock Holmes’
Over the years, MyHeritage has benefited from a lot of exposure in Israel and abroad, and launched unique initiatives such as documenting family histories and cultural heritage of tribes at risk of disappearing. From 2014 to 2019 it led a project for documenting gravestones in cemeteries across Israel, assisted by tens of thousands of volunteers. In 2016, the company distributed free DNA testing kits to anyone who might have some connection to the missing Yemenite children affair, which led to a few emotional reunions.
MyHeritage has also launched unique initiatives such as documenting family histories and cultural heritage of tribes across the world at risk of disappearing
Japhet received the moniker of an Israeli Sherlock Holmes from a French journalist, after taking on as a hobby the location of legal heirs to the owners of art work robbed by the Nazis and held by French museums. He managed to trace eight heirs and restored 20 works of art to them. At the same time, the company was also subject to criticism and hit several potholes.
One of the company’s main growth engines in 2017-2018, DNA kits for genealogical testing, started winding down. A new product it had been working on for two years, medical DNA kits, turned out to be a wasted investment. Their weakness was experienced by the company’s competitors as well. The main explanation for this is that the market is saturated and limited in size. Most of the people who wanted to take a genetic test have already done so. Moreover, there are no repeat users. Anyone who’s done a test is unlikely to take another one. Consumers also discovered that the value of DNA tests is rather limited. Insights on ethnic origins are usually only curiosity items. Medical information includes statistics regarding an increased chance of contracting one disease or another, and people obtaining this information don’t always have use for it.
However, the drop in interest has an additional cause: privacy concerns. Consumers are starting to worry about the future of databases containing their DNA. In addition to concerns such as whom this information may be sold to and with whom will it be shared, there are always concerns about leaks or data theft.
DNA information uploaded by citizens to open databases has assisted the police in their investigations in recent years. In a particularly famous case, a man was arrested in 2018 on suspicion of being the Golden State Killer from the 1970s, after the police used a DNA sample put on an open database by one of his relatives. Investigators examined the family tree on the website GEDmatch in order to mark persons matching the suspect’s profile, and then tracked him down. This case increased fears that governments and law enforcement agencies will misuse such information.
"if I could go back in time, I’d erase this medical adventure. It wasn’t successful and also wasted a lot of resources. It did help us by giving us the confidence to set up the coronavirus lab, which I’m very proud of. But medical DNA was a failure"
In the field of medical DNA kits, MyHeritage’s competitors, first and foremost 23andMe, have adopted a new, lucrative business model of selling genetic databases to pharmaceutical companies for use in developing drugs.
Japhet was attracted to medical DNA kits out of curiosity and a belief that this would help him attract a new audience interested in exploring their roots.
“We inherit our genes from our parents and from their parents,” he said upon launching the product. “It’s very important, when you discover a potential genetic or medical problem, to understand which ancestors you inherited the problem from and how severely they developed it.”
He promised that family trees and DNA data are protected in separate systems with additional layers of security, that everything is stored anonymously, and that the company takes extreme measures so that even its employees don’t have access to users’ medical data.
“The very fact that we offered medical DNA tests hurt our reputation,” he said. “Many European countries started to closely examine us because of it.”
He admitted that, “if I could go back in time, I’d erase this medical adventure. It wasn’t successful and also wasted a lot of resources. It did help us by giving us the confidence to set up the coronavirus lab, which I’m very proud of. But medical DNA was a failure.”
Users’ suspicions also put MyHeritage in an uncomfortable position after it hit another pothole. In 2018, it suffered a security breach that led to the email addresses of 92 million users in its database being exposed, along with hashed passwords.
The damage to users wasn’t as great as the harm to the company’s reputation, and Japhet still refers to this as a traumatic event. For three months he slept at the office, ordered all projects and development work halted and mobilized all the company’s engineers to improve security.
Currently, 920,000 users pay annual subscription fees for MyHeritage’s services. The company’s total bookings, including from sales of home DNA testing kits and coronavirus testing services, are estimated at $268 million in 2020. MyHeritage has been profitable since 2012
One of every five Norwegians
Japhet believes MyHeritage is still far from realizing its core service potential.
“In some countries, especially Scandinavia, we’ve hit a glass ceiling. For example, in Norway, whose population is smaller than Israel’s, one of five people is a registered MyHeritage user, and more than 1 percent of the residents are paying subscribers. This amazing subscription rate means that if you’re Norwegian and register to the site, you can give me very little information and I can tell you a great deal, because of the information your relatives have provided. We’re approaching a similarly high coverage in Denmark and Sweden as well, and we’re moving toward it in The Netherlands.”
But Billions of people worldwide still haven’t heard of MyHeritage and the company has a lot more to do for them, especially in the United States, Canada, South America and Europe, for example in giving them access to historical documents, he says. Over the years MyHeritage has purchased companies that operate historical information sites and is constantly creating collaborations to enhance its data repositories. It is capable of matching documents to relevant family trees in 42 languages.
The focus on family roots characterizes relatively wealthy countries. “In Africa, people are less interested in how their grandfather lived and focus more on how they will live. Digging into the past may improve the future, but to attend to it you need the present to be fine,” Japhet says.
Japhet likens family history research to carving the Siloam Tunnel under Jerusalem, to provide water to the city some 2,700 years ago.
“According to the Siloam inscription found in the tunnel, the tunnelers met mid-way in some kind of miracle,” says Japhet. “Exactly in the same way, when a genetic match is found on our site between two users, each researches from their own direction until they find the meeting point – a common ancestor.”
“Searching for family roots is extremely viral, you’re exposed to a product feature or a document that makes you want to enter the site and add information of your own. This information in turn serves all your relatives who are registered on the site and ‘awakens’ them from their slumber.”
“We’re a company with subscribers and information whose value appreciates with time, unlike the rest of the internet. The major part of our income comes from recurring subscriptions. It’s a robust business model, because it means a stable income forecast. The reason for our high customer retention rate is that people see their family history as an asset they want to keep for life and pass on to the next generations. The percentage of renewed subscriptions on MyHeritage for subscribers of more than three years has been close to 95 percent They already know they’ll stay with us for life,” he says.
From COVID to home AIDS testing
MyHeritage set up its coronavirus lab last May as a humanitarian project aimed at helping Israel fight the pandemic. Since its establishment, the lab has carried out 3.25 million tests.
Sixty contractors and 200 construction workers and electricians labored to build the lab in a record three weeks in the Kodak building in Petah Tikva. The lab’s ceiling was constructed of hundreds of steel panels covering an area of 1,200 square meters so that it could hold the air conditioning tubing, whose total length is about three kilometers. This enabled the lab to maintain the requisite safety conditions – low air pressure and replacing the air 26 times an hour.
When Francisco Partners priced the company for acquisition, it decided to assign zero valuation to the lab, because its future was unclear
The laboratory equipment isn’t connected to the internet, and it’s completely separate from MyHeritage’s servers and the rest of the company’s infrastructure. This was the first lab in Israel to be built from scratch for the sole purpose of doing coronavirus tests. MyHeritage invested $5 million in it, and a similar sum came from its Chinese partner, BGI.
Japhet said that when Francisco Partners priced the company for acquisition, it decided to assign zero valuation to the lab, because its future was unclear. Francisco did end up buying the lab’s activity as well, but Japhet said this was a strategic decision on his part.
“The lab was set up to meet a humanitarian need, and I didn’t want someone with a financial interest to come and ask why we’re maintaining it,” he explained. “So I added a provision to the contract that as long as the lab is contributing to Israel’s battle against the coronavirus, they are forbidden to shut it down.”
But it now looks as if the lab’s role is nearing its end. In the past, MyHeritage carried out around a third of Israel’s coronavirus tests, but today, it performs only about an eighth of them, because other private labs have opened and reduced the country’s dependence on MyHeritage. Moreover, incidence of the virus has declined significantly and the number of daily tests is falling steadily.
Over the past few months, the lab has fired about half of its employees. Japhet attributed some of these layoffs to improvements in the lab itself. For instance, greater use of robots and automation reduced the need for workers.
The lab also started making greater use of pooling, in which several samples are tested together, and only if the pooled sample produces a positive result are the samples comprising it retested individually. This reduces the number of tests that have to be performed.
"Swab tests using the same equipment and staff we already have in the lab could also test for AIDS, Ebola, rabies, papilloma, chlamydia and gonorrhea"
Japhet still employs 80 people in the lab and promises to put it at the state’s disposal as long as it is needed. But he’s already examining new lines of business that would utilize the expertise, equipment and staff he has acquired.
“The qPCR equipment [which is used to amplify DNA segments ] is good for detecting infectious viruses, and the lab’s high level of biological safety is also suited for this,” he said. “The coronavirus is an RNA virus, and there are other viruses of this type, like AIDS, Ebola and rabies. Thus, swab tests using the same equipment and staff we already have in the lab could also test for these diseases, as well as for papilloma, chlamydia and gonorrhea.
“Home testing for sexually transmitted diseases has soared worldwide, because people prefer getting tested at home, privately, to going to a hospital or clinic,” he added. “You buy [the tests] from home, send them back to the lab and get your diagnosis privately. These tests wouldn’t serve the Israeli market, but the European and American ones.”
If any such decision is made, Japhet said, business logic would dictate making the lab an independent spinoff rather than keeping it as part of MyHeritage.