When I did my own research and analysis, most advice I found online was that they weren't horrible; that private policies were preferable but employer-sponsored policies were okay. But IMO I really think that understates the cost-effectiveness in a way that misdirects people. That advice provides an excuse to just buy the employer-sponsored policy and move on.
Without being too long-winded, if you just consider the underlying economics of insurance, you'll get the best deal by purchasing and maintaining a policy for a longer period and with a payout that best matches your current and expected salary. If you're healthy, relatively young (That last aspect is key--maintaining the policy for the entire term. Your risk increases significantly as you get older. If you don't maintain the policy for the entire term, you've basically thrown most of your money away. The fact that people are very likely to change jobs multiple times in middle-age (especially programmers!) means that you're not only introducing gaps in coverage, but increasing the likelihood that you either fail to buy a new employer-sponsored policy, have an employer that has a worse policy, or one that has no policy at all. That to my mind is a huge downside; difficult to quantify but so significant I don't think we need bother precisely quantifying.
Also, only small polices (One final tidbit: employer-sponsored premiums are pre-tax, while private policies aren't. OTOH, payouts on the former are taxable, while (at least for term life) the latter are tax exempt. Normally pre-tax is a better deal, like with health insurance and retirement funds. And it _could_ be, theoretically. But it's not. A $1m employer-sponsored policy is really only going to payout about $750,000 or less, but the premiums aren't discounted to match even considering the pre-tax advantage. I don't remember how much time I spent crunching the numbers--most of my calculations were back-of-the-envelope--but I'd be surprised if my conclusions were wrong in this regard.
Basically, just buy a private policy. I purchased _less_ than policygenius.com (and standard financial advice) recommends based on my current compensation because I didn't want to end up in a situation where I allowed the policy to lapse because of the premiums. Again, that's just throwing money away. At some point you just have to pull the trigger because you can become paralyzed thinking about this too much. I probably erred to low, but I'm absolutely sure I didn't err in choosing to buy a private policy.
And remember, if the issuer isn't sending someone out to your house to draw blood[1], your policy is much, much too small. The questionnaire you fill-out for policygenius.com or anywhere isn't for naught, especially the compensation answers. IIRC, I had to send in proof of my compensation after I chose a policy and put things in motion. Unless you're actually getting a poverty wage, any cost-effective policy will require a health check, and the best premium/payout ratio will be in the range of something like 10x-15x your current annual salary, rounded up or down to match the nearest, most standard policy amount.
[1] They sent someone to me. At the time my employer-sponsored policy required me to go to a doctor. I don't know which is more common, but FWIW at least some issuers send someone to your door[2], which is very convenient. When I initially joined my then employer I did elect their sponsored policy. But I never got around to going to a doctor for the required health check so it never activated.
[2] The lady was nice and her job seemed like a pretty cool gig for a phlebotomist.