How High-Flying Zenefits Fell to Earth
buzzfeed.com
buzzfeed.com
Excerpt:
Zenefits’ technology was frustratingly limited. The software is built around a central dashboard and lets small businesses onboard new employees, generate health insurance quotes, and manage 401(k) plans and other benefits. But while this system looked slick and did a fine job with small tasks, much of the important work on the backend was done manually.
“The back office is really heavy in terms of bodies,” one former Zenefits employee said. “It’s really archaic. It just seems automatic to the customer.”
This kind of work is standard practice in the insurance brokerage business, but it’s notable for Zenefits, which purports to be an “automated” solution. The lack of true automation behind the scenes led to numerous problems. With one customer, Plated, a food delivery startup in New York, mistakes by Zenefits staff in entering payroll deduction information caused a string of erroneous paychecks to be issued, according to one former employee and an online comment by Plated co-founder Nick Taranto. Zenefits is “not connected well enough with payroll providers,” Taranto wrote, “so we’re having to cut lots of manual checks.”
The first public indication that something was amiss at Zenefits came last summer, when BuzzFeed News reported that customers were complaining of software glitches and seemingly careless errors.
ADP, Trinet, etc. are just as manual and full of error prone human processes. It's all just heroic human effort on all sides. If anything Zenefits has shown the incumbents and everyone else in the space that you can indeed provide a software solution that integrates well with human processes and in the long run make things much more automated and efficient.
I could see how a company could try to fake it until they make it- as in, until they get enough clout (via money, customer base, important friends) that the companies they're trying to automate realize that it is in their best interest to work with them. However, while trying to make it, I would think that you would do everything humanly possible not to screw up even one piece of data for even the lowliest of customers. If you are 100% above board and have a really good process in place, when things happen you can be pretty sure it wasn't you, and you can be in a better position to actually help the client.
FFS, Zenefits is screwing up the only thing about employment that matters to a lot (most?) people: the freaking paycheck and the health care. Hearing all these horror stories it becomes hard to see how Zenefits gave a crap doing their job at all- I can understand a mistake, but a mistake with a devil may care attitude means you've lost my business forever.
I've had a little exposure to health insurance and the medical bureaucracy and I've worked as a government contractor, and in my experience the healthcare world is far more sclerotic than the Federal bureaucracy in many areas. The Feds are also much more open to automation and newer technology, while in medicine they literally do everything with paper files and fax machines and will not budge.
Arroyo, who runs the Los Angeles software agency, made Zenefits his company’s health insurance broker in 2014. He experienced a string of minor problems at first, he said — annoyances like an inability by Zenefits reps to explain a mysterious charge on his insurance bill. Then, last fall, he was told he would have to submit to a random audit by his insurance carrier. He gave his information to Zenefits, trusting that it would make its way to the carrier.
“A week or two later, I got a letter saying I wasn’t compliant, and our insurance had been canceled,” he said. Zenefits, he said, “never gave me an answer of what happened. They just said, ‘Oh, it must have gotten lost somewhere.’” Arroyo dropped Zenefits in December.
Go around and ask people how much they like ADP and Trinet. Chances are you'll find more than a few customers that hate them with a passion. Not just because of clerical errors like the ones I mentioned but because of how bloated and error prone their products are. I recently had to change my address in the ADP online portal and I couldn't do it. I had to reach out to both their customer support and my previous employer to get them to change a single text field.
Maybe Zenefits lacked imagination? You can automate sending things via fax machine. OCR exists. You can use web scraping tech (well that's kinda what got them in trouble with ADP right?)
I feel like it's always possible to automate a bit more (or at least enough to where you don't have staff manually copying numbers onto a piece of paper)
I've spent a fair amount of time working on infrastructure and operations projects for insurance MGAs (a particular kind of broker), and can assure you that these approaches are definitely in use by many players within this space. If Zenefits hasn't implemented these methods, it would be pretty unusual.
Granted, even in spite of heavy automation in the broker space, there's still a ton of very old school paper shuffling and processes that resist automation because of a combination of the following factors:
* lack of standards * process complexity * the sheer number of 'one off' cases that pop up in the insurance world * when you do the cost analysis, it's often cheaper to rely on a manual process with manual controls than it is to automate, thanks to relatively inexpensive clerical labor
Personal anecdote: Enrolled for health insurance through the Zenefits portal. Waited a few weeks for my insurance information in the mail. Eventually I got an insurance card for someone with a different name and gender, but with my address, phone number and even social security number! I even got the insurance nurse hotline calling me for this person, but not for me. It took Zenefits 2+ months to correct this data entry issue.
Another anecdote: The insurance plans we had selected with our account rep were different from the ones offered through the Zenefits portal. We assumed these were the same, but employees were actually enrolling for worse plans than we had thought.
Account managers were generally very unresponsive, then again - we were just a small startup (but with very simple, standard needs).
That said, a previous startup that did not use zenefits fucked up DOBs for both me and my partner, causing her to be unable to fill scrips for a while. So they're not the only ones with issues.
On June 4 (2015), we disabled Zenefits access to ADP’s RUN small business solution due to unusual and alarming demand for data from Zenefits far out of proportion to the number of clients who have allowed them access to our system. Despite Zenefits serving less than 0.25% of the clients on our system, they had been responsible for up to 25.0% of the total user traffic (in other words, a hundred-fold times ordinary user traffic).
…it was pulling sensitive information, including unmasked Social Security numbers and employee banking information, in a manner that did not comply with ADP’s standards for data security.
http://www.businessinsider.com/adp-and-zenefits-are-at-war-2...
The app was just a front-end. When customers ordered something, people sitting in call-centers would manually order the items at restaurants.
There was zero compliance, widespread scams and no checks and balances, all along with minimal technology.
Full story for anyone interested: http://www.livemint.com/Companies/rYKC6HjnShogjE62jO5lpK/The...
I'm becoming more and more cynical of the startup world and the press that makes founders out to be heroes.
TopTable was eventually acquired by OpenTable (losing lots of its charm in the process).
Seems like the problem here was either choosing not to or simply not being able to automate their process.
> “It’s almost as if they want us all to quit so they can start over.”
Oh, wow, gee, that's so awful! Wow! I'm so sorry you can't get away with lying to the government and your customers any longer. I'm so sorry you can't get drunk at work and pretend you're still in college! Oh, this is just awful. Won't somebody do something?
(/s, as if it were necessary)
The "move fast, break things, grow as much as possible" mindset is not always a good thing. If the product you're making has high potential and low risk, if it's replacing something that isn't critical to people's safety and livelihood or to your customer's day-to-day operations, it's great. If you're making something new the world has never seen before, it's great.
But if you're making sure people are getting paid and have medical coverage? That's not an area where it's an acceptable mentality. I think the issue here is not so much that this ethos is fundamentally flawed (though I guess one could make that case), it's that it is not universal and shouldn't have been treated as if it were.
Also - respect to Buzzfeed for taking what could have been a hit piece and making it fairly comprehensive/thoughtful.
(edit: formatting)
Deleted comment
Yes, thank you. Keep it up please.
In an ideal world, people/employers should just be able to do it themselves and cut out the middleman.
This would not have happened if the person selling you the insurance was under a legal (read: license) obligation to provide you the most appropriate insurance.
I could see how even an old-school agent could forget to mention earthquake insurance for home insurance or lack of acupuncture in health plan.
And many, many, many more. By comparison, advising on health insurance actually kinda makes sense.
California licensee lookup: http://www2.dca.ca.gov/pls/wllpub/wllquery$.startup
When the vast majority of customers can't tell (even after years) whether they've been had, that creates an unusually strong argument for having someone else check whether the seller is being honest, such as a licensing authority.
(Which is an argument for mandatory auditing, not for licensing. Yes.)
In an ideal world everyone would do the right thing, in the real world some level of bureaucracy and regulation is a necessity.
A bad hairdresser can slice off my ear and a bad cook can poison me.
In finance, it's incredibly easy to make a living by acting in bad faith. You can offer insurance with no serious intention to pay out, and tie up claims in bureaucracy and legal loopholes.
Good regulation should go some way to minimising bad faith actors.
In fact the industry seems quite scammy anyway. But it would be much, much more scammy - to a criminal extent - with no regulation at all.
Both of which require licensing.
If I flip your concern around, why do you think it would be appropriate for someone to sell me a health insurance plan without any understanding of what the different plans offer or which plan best served my needs?
If I'm buying insurance, I expect the salesperson to be able to understand my situation, what I'm looking for, and tell me all my available options, and the benefits/drawbacks of each.
Licensing makes sure the salesperson goes through the efforts to get to this point.
I signed up with Covered California a few months back and while they provided basic facts and a comparison matrix of plans, they didn't go out of their way to make sure it would fit my needs just perfectly. With nothing but a Web site and a wizard, how would they?
If so, it sounds like insurance salespeople are similar to tax or financial advisors, and regulation forces them to read through and familiarize themselves with all of the options. Fair enough. Just wish the system itself was revamped, pricing schemes simplified, and information made more accessible, to make even this unnecessary.
One mans bug is another mans profit. So that's not happening.
It's a legal mechanism that stops insurance companies selling people policies that don't actually cover them for what they need. Essentially it shifts the understanding of what insurance is required from the buyer to the insurer. It's harder to refuse a pay out if your salesperson has to make sure all the necessary boxes are ticked.
Well, reading what Zenefits was doing, and how badly they screwed up might enlighten you.
Another reason: there's a reason startup offices are so "fun": they don't want you to leave. Nerf guns and hard liquor make the office seem like a place to hang out at, so working 10,12,14+ hours a day doesn't seem so bad. You're hanging out with your friends, right? Overtime pay? Never heard of it. We get stock options right? No, our founders wouldn't dilute our shares to favor late stage investors, would they?
So maybe that's the problem.
Apple really championed the "beer bust" culture in the early 90's.
I have been in "dry offices" and do find it a morale killer, because there's the sense that the HR department knows better than you.
We have beer in our office, and a bottle or two of hard stuff, but honestly no one ever touches it. We have a hard working and professional culture and know alcohol and coding don't mix. ( Or, as I say, "Don't drink and write drivers." )
Sometimes the inside sales guys knock off at 3pm on a friday and have a can of beer. They start early - calling the east coast - so it's good to see them blowing off a little. More professionally, we tend to troop to the dive bar around the corner, which is honestly more fun than setting up a kegerator.
Having beer and not drinking it is a more solid culture than banning beer from the top.
And it has consequences.
The startup culture is very toxic in terms of use of psychotrops. People takes all kind of drugs (alcohol included) often with a laxist attitude from supervisors and drugs + workload + stress = lot of catastrophs. Ho! And of course, the more crunches the more open bar there are to reward you from keeping on with the good job and release the pressure. EDIT and it happens that drinking overlaps with production duties during crunches, when people are already lacking of rest. Thus, generating of course more production incidents.
Startup are brain crunchers and are free to go on since there is no legal liability for this, and they don't pay the price for the damages since they just fire burnt out employees who have to pay for fixing themselves afterward and tell customers there are no guarantees software works, it is a lose-lose-win situation.
Didn't see the previous ones, [0] and lots of them. What did stick out, was this quote:
"“Why are you guys so fucking bush league?” Dalgaard said, according to Conrad, who recounted the conversation while on stage at a software conference in February 2015. While Conrad had planned to hire 20 sales reps by the end of 2014, Dalgaard said the target should be five times that. “You guys gotta get your heads out of your asses, start focusing on going big here,” Dalgaard said, according to Conrad."
The expectation from Lars Dalgaard of Andreessen Horowitz to grow big at ^all costs^. VC money egging on the unethical.
[0] https://hn.algolia.com/?query=zenifits&sort=byPopularity&pre...
Hmmmm, "a person familiar with the matter" sure makes it sound like Dalgaard was doing everything in his power to encourage Conrad to act responsibly. I wonder who this person could possibly be and whether they have a personal agenda of some sort.
Does it strike anyone else as maybe a little bit irresponsible that a partner at one of the world's preeminent VC firms would pressure a founder-CEO to hire as aggressively as possible and shortly thereafter castigate said founder-CEO for their company exhibiting all the telltale signs of hiring too quickly?
As a technical founder of my own company, the founder of Zenefits ( reportedly, they maybe use that line but coming from standard sales reps ) reached out and suggested I take a look at Zenefits.
I looked. I asked two different "startup CFOs" about Zenefits.
I found the following conclusions:
* The market is crowded. Trinet is what everyone in the valley has been using, and no one has serious problems with them. It's not like, you know, fighting the intrenched taxi mafia. Zenefits is not disrupting anything, they're simply competing.
* According to my sources, Zenefits is the hardest company to work with of benefits suppliers. Compared to Trinet, when your CFO-type calls Zenefits, they get long wait times and unprofessional answers. One of my sources was forced to use Zenefits ( by a founder ) and considered them a time sink.
* High prices. Because of a small installed base, Zenefits didn't have the ability to create a large insurance pool and have market power with insurers. They claimed low prices, and did have low prices for their service, but the numbers penciled out higher because of higher health insurance prices. If they think prices go down when they get bigger, maybe the would have needed to subsidize prices.
Not a win.
Part of our startup culture these days is "damn the law, make customers happy". I'm not super thrilled about that attitude, but it is what it is.
The zing against Zenefits in this article is poor technology ( uh, chasing MPV is all about poor technology ), and poor legal practices. I honestly didn't care about either - I care about user experience and prices.
My internal recommendation not to use Zenefits was poor user experience, and high prices. I didn't like having a sales person tell me prices were low when I had done my research and prices were higher. I didn't like being told Zenefits was easy to use when my people said it was harder to use.
And, yes, I sent a note outlining all the to the founder, in the interest of helping a brother out. I don't remember what I got back, but it wasn't encouraging.
I don't get why the passing grade is so low. At the very least, I'd want someone who only scores a 60 (and thus is barely qualified) to be forced to take it again the following year to ensure they're still qualified.
I wonder if you could enforce a graduated system based on scores? If you score a 60-70, you have to re-certify the following year; if you score a 70-80, you have to re-certify in two years, and so on. In that case, scoring as high as possible is in the best interest of companies, because then their employees go as far as possible before having to retake the test.
Sounds like an awesome place to work.
Problems such as being without proper health insurance can cost lives.
I myself have been forgiving in the past about such things; but that attitude was born out of an ignorance I had that everything on the back end was being handled as well as possible. After reading this article I am now much more concerned.