How enterprise software is like baby clothing | Hacker News Reader

How enterprise software is like baby clothing

twitter.com

84 pointscs70266 comments

https://threadreaderapp.com/thread/1182635589604171776.html

I work in K-12 EdTech and this hurts to read. Every sprint spent working on a "check the box feature" that administrators want to see is time not spent improving the effectiveness and usability of the product for teachers and students. Worse, sometimes the features or changes administrators think they want actively work against the interests of the users.

Edit: And god forbid you end up working on a core curriculum product that needs to be reviewed by each state's board of education. In that case you end up building a product that appeals to state-level bureaucrats, just to get get the privilege of selling to administrators, with little thought left for the actual students.

Oh I thought this was going to be along the lines of Michael Dell's assertion. He said ever time Dell doubled in size, he needed to revisit all the processes. Automate things that used to be manageable by hand. Customize processes that used to be one-size-fits-all. Because a bigger company has different optimal solution points for nearly everything.

(Which could be taken to heart by people insisting that government should go back to some imagined simpler past. Those old processes are obsolete; there's no going back. But that's another tale.)

I'm curious in cases like this if it's worth growing so much as a single company. It feels more efficient, in my mind, if the company splits up in smaller pieces everytime that demand grows and products change. It's just a feeling.

I don’t have a link but go read about Gore, Inc (of gore-tex fame) which did/does something along these lines.

It's relevant that more people in software tend not to understand this either, despite our entire profession being rife with these sorts of problems.

The non-linear scaling of sorting algorithms is a classic one, and even the different optimization points is baked into the "optimal" solution - i.e. quicksort when your number of elements is below a threshold is better replaced with an insertion-sort, and indeed this is what most implementations do.

Which is an algorithmic example of solution overhead dwarfing problem size (and yet if you're above that threshold, the more complex solution will always win).

> Which could be taken to heart by people insisting that government should go back to some imagined simpler past. Those old processes are obsolete; there's no going back. But that's another tale.

It isn't imagined, and there's no reason government ever needed to turn into a giant, bloated bureaucracy that's more interested in sustaining its own existence and growing ever-larger. In the United States, for example, the federal government alone is the largest employer in the entire world. Yes, bigger than the governments of countries that have three times as many people.

Loved the rest of your comment though :)

A giant country needs more governing. Dreaming about 'going back' is fantasy.

Sure there's plenty of things to improve. Just don't break everything before fixing it; people's lives depend on it.

> a giant, bloated bureaucracy [...] In the United States, for example, the federal government alone is the largest employer in the entire world.

That's misleading: You are asserting "lots of pencil-pushing bureaucrats", but the numbers you're touting instead rest upon "lots of gun-holding soldiers."

Depending on how you crunch the numbers, India or China may have taken the #1 spot by now.

P.S.: In contrast, consider the OECD stats for "Employment in general government as a percentage of total employment" [0], where the USA ranked somewhere like *squints* ~22nd in 2019, below the OECD average.

[0] https://www.oecd-ilibrary.org/sites/75f92d89-en/index.html?i... , graph 3.1

The supermajority of employees of the federal government are from military/defense growth explosion sustained since WWII. The US is also not top of the list these days, India's department of defense has taken the crown. I.e. being a top employer is and has been mostly due to having a massive military, not bureaucratic growth, even outside the US. It's also worth noting federal employment numbers peaked 30 years ago, when there were 25% less people living in the country, so it's by no means been continuously growing ever larger.

https://usafacts.org/articles/how-many-people-work-for-the-f...

A super majority is a specified proportion that is in excess of a simple majority (e.g., where a simple majority would be 51%, a super majority might be defined as 60%). Per your link (which is admittedly interesting), military/defense only accounts for about for 770k of 2.2M full time federal employees. That is not a super majority, that is a plurality (the single largest group but not a majority).

The supermajority is in reference to both military/defense, though I could see how that could be very confusing wording with the "military-defense" category also on that site. The difference being non-military level defense numbers were not included in your calculation of the pularality. Non military defense also at least includes Department of Homeland Security, for example. Between those two you've already left plurality for majority. To get to the supermajority you need to bring in the department of veterans affairs. Note these 3 compose the top 3 offices in terms of employment.

Now I suppose you could argue the latter as bureaucracy as it's not direct defense but even then it's sole purpose is to provide lifelong healthcare support and benefits for the military personnel figure, not just random bureaucratic growth. Since that site didn't actually categorize the types of office, this reference includes similar defense categories as I listed above (though it also includes the Department of Justice, so gets a slightly higher percentage than my mind at the low 60% area): pdf warning https://ourpublicservice.org/wp-content/uploads/2022/03/FedF...

Which could be taken to heart by people insisting that government should go back to some imagined simpler past. Those old processes are obsolete; there's no going back. But that's another tale.

We definitely could create smaller governments to slow the consolidation of power. Splitting the more populous states in half or whatever, and treating the remaining states more like the EU treats it’s states would be a huge step in the right direction. but as you said it’s another tale.

Why would adding more moving parts help? Standardization is a good thing, and America is very bad at it if you ever try to do anything that crosses state borders.

Because "machine" is a poor analog for government activities.

The goal is not efficiency, the goal is to minimize the amount of damage that can be done by a single person or group with evil intentions.

On that vein, we could also say that if Elon Musk was capable of firing 80% of Twitter and keep the company performing the same functions, maybe the same could be done to governments.

Twitter today could hardly be described as performing the same functions as pre-Elon.

The fact that the entire thread is not readable without an outside app or an account is but one of many things that has changed.

Half of the time, I wasn't able to read threads (or even the linked tweet) without an account even before Musk's acquisition.

Explains how ServiceNow can be so popular yet so awful. They sell to the IT department who staffs up a team of solution architects and lock themselves in.

Usability of the software be dammed (it took years before even urls were shareable).

Slack was a revelation where its pricing model encouraged bottoms up adoption (only pay for seats used) and they encourage feedback from everyone (unlike Google docs where it takes an admin to report a bug, maybe).

>where it takes an admin to report a bug

Unfortunately most products seem to conflate "report a bug" with _enterprise support_.

If you think ServiceNow is awful, then you clearly haven't used the alternatives.

Believe it or not, there are users (not buyers) that _love_ ServiceNow for what it can do. Usually because what came before was a mixture of manual process, spreadsheet hell, or some living nightmare like BMC Remedy (at least for ITSM use cases).

ServiceNow is great if your project/architecture people are really great. The mistake is expecting that OOTB features will be enough (with only a few tweaks) or that the product itself will steer you toward best pratices (NOPE!). Its a bag of great parts which require extremely careful assembly and a disciplined approach in implementing the simplest, and most elegant build. YAGNI. Case>Req>RITM>SCTASK is a lot of layers of bloat just for a single user request.

I've heard this same thing said about Workday. Great for HR, terrible for everyone else. it's wild to me that my company uses it and doesn't find it insane that they need a multi-step "how to" doc on how to find your own performance review, one of the 3 things that the software is used for.

At least with the workday, you aren’t in it all the time, but yes, it reminds me of phpmyadmin over a database that needs a 5 hour maintenance window every week.

Ha! Workday, datadog, salesforce, Odoo, aws... etc

All software with huge amounts of functionality suffers from this. And some of them even charge you extra for their "learning programmes".

As a parent I can say the magnet clothes is amusing and interesting and insanely expensive and not worth it.

Because just like enterprise software, the users are just going to shit all over it anyway.

I wouldn’t say they’re insanely expensive - we got 2 magnet sleep suits for $40. They’re super soft and really easy to put on the baby. Check out consignment stores specializing in baby/kids clothing for a good deal, or just buy new for $20-$40 each. So far, no blow-outs.

How bad are those in the washer / dryer and how soon does the magnet abuse kill the magnets?

Feels like this is a pointless techy solution to problems already solved by Velcro. I don't have kids, am I wrong?

Velcro is ok but snaps that rent too strong/tight are just as easy. Velcro can “misalign” if it’s too long.

Velcro is actually substantially more annoying to launder because it can clings to fabric.

If you wear pajamas to bed, imagine all button holes on the pajamas were replaced by Velcro. Now imagine the sound and comfort of the Velcro as you roll over at night.

They would stick to the side of the dryer sometimes but no worse for wear, I saved 'em for the next baby.

Quite a lot of my friends here in Sweden who got kids lately ended up just buying tons of second hand velcro-closed baby clothes, paid by weight.

I don't understand why that isn't a much more common practice across the world, it doesn't make sense to me that people are buying new clothes for kids that will outgrow them in 1-6 months time.

The clothes are sold to the parents, not the kids. But secondhand shops and similar exist the world over.

Wow totally disagree, I loved the magnets during the pre-crawling / houseplant phase. So convenient for changes.

Concur fits this analogy well. I'm sure it was great for the accounting team but using it to do expenses (at least couple years ago) was painful. Having used Brex and Ramp they both seem to be the Magnetic Me baby clothes versions of expenses

The baby clothing analogy doesn't quite fit because the people buying the baby clothes as a gift are in a relatively close position to talk to the parents (i.e. the end user). They just didn't do the work. The buyer/end user seperation in enterprise has more layers of separation (it can be an entirely different department) and stickiness (you are forced to use the terrible baby clothes).

I wouldn’t dismiss it that quickly: I think it works when you consider the kind of buyers like grandparents who are decades removed from the day to day work but still consider themselves experts. I’m also reminded of a few executives by comparison to grandfathers who assumed it wasn’t that hard because their wives did it.

The administrators are also in a relatively close position to talk to the users. They share a campus with them. They just didn't do the work.

The analogy holds.

I used to work for a company producing E-Learning software, and also content.

The customers were never people who wanted to learn something, but companies who wanted their employees to learn something, and often companies who had to comply by government mandates about making sure their employees were aware of some things, like money laundering laws for instance.

In my observation the companies were not really interested in teaching their employees how to prevent money-laundering, but how to prove to the government that they (the bank etc.) were making a serous effort to combat it by training their employees.

I hate most simplistic metaphors but this one is insightful.

This is a big reason why Microsoft blew the mobile phone biz: their customers were not the users; the customers are central IT. The iPhone was such a revolution at the time that it was enough to force IT to allow the alternative. And from that MS had no feedback system to learn what they were doing wrong.

This is true.

I have a niche e-commerce in this space. People loved it so much that the app went down for a few seconds and then we got A LOT of angry calls!

The app is made for on-the-street salesman so it is optimized for fast invoicing.

To avoid the problem that the manager's needs for reporting and dashboards overshadow everything else we split that from the main app. i.e: we have no report capabilities in the transactional. app interface and have instead integrated metabase for the managers.

So, we split OLTP/OLAP as ux measure.

I am not sure I got the baby clothing analogy, but I support the statement that Blackboard is pretty bad. It works, but for many things you need an instruction telling you exactly what to do as otherwise it's very difficult to find what to do/do the right thing, given how many menus/options there are. On the other hand having experience with Canvas, that was pretty good and painless.

How things change - I remember when the university I'd previously worked at finally managed to kick IBM out and replace the universally reviled Lotus Learning Space with... Blackboard! I guess it metastasized over the years and is now the big bad one that gets replaced by something else.

He makes what I assume is an estimate that Bb is “20 years old” and I find that amusing because I remember using it in 2003 and it was definitely already creaky and old then!

On a side note, you do not want the convenience of baby-clothes equipped with magnets for easy-on easy-off. Baby's also got an easy-to-the-mouth to consider...

Anyone else get flashbacks of evaluating software purchases for a company, only to be completely overruled by someone who was clearly getting some kind of kickback from the vendor?

I kinda wish someone would set up a class action lawsuit for all the stress caused by having to deal with Jira, HPSM, and the like.

Oh, you’re counting clicks. That tells me all I need to know about the author’s UX knowledge.

Some things do need friction. Just like with baby clothes. Something too easy to remove will leave you with a perpetually naked baby.

And that’s what the tweet thread is about, she doesn’t like Blackboard’s UX. And that’s fair, it might not suit her particular needs. However, a bespoke solution is going to cost far more than any number of Blackboard licenses. And Blackboard is going to gun for “mostly ok” for “most teachers”.

It’s essentially the Word paradox. Everyone only uses 10% of its features, but everyone uses a different 10%. You can’t optimize for that.

I used Blackboard for a decade. Trust me, it's awful.

As a student who used blackboard in college it was terrible

Have you used Blackboard? Did you feel it was so full of features that you only used 10% but the parts you used were solid? I remember disliking it thoroughly.

I think the other posters have correctly identified the problem as the principal-agent dilemma.

Blackboard could be awful, but not for the reason she states.

That’s the issue here. Her giving UX advice would be like software developers making education software without input from educators.

Minimizing the number of clicks -- work required to accomplish a task -- is actually a valid focus of UX, called information architecture. It even has a pithy rule: https://en.wikipedia.org/wiki/Three-click_rule

I’m not trying to take a stance in this disagreement but your own links says,

> The three click rule has been challenged by usability test results, which have shown that the number of clicks needed to access the desired information affects neither user satisfaction, nor success rate.

That's a bad summary and defines common sense, as stated. The situation where more clicks won't lead to dissatisfaction is when results are progressively disclosed. If you can design your UI to achieve the same thing with less work and without overwhelming the user you should.

https://www.interaction-design.org/literature/topics/progres...

The guy is a CS prof at Princeton. He worked with Mozilla on Do Not Track in HTTP headers. It's not some random teacher. I think he has a little idea of what UX should at least look like.

https://en.wikipedia.org/wiki/Arvind_Narayanan

He might know it is important, but if you look at his CV, it's mostly non-user facing issues.

And that's not to diminish what he's done. I don't expect everyone to be an expert in all fields. I wouldn't get someone who only wrote microcontrollers to redesign my frontend. And I wouldn't get a CS professor to write my microcontroller software.

I wouldn't go to GP for cancer and I wouldn't go to a Oncologist for physical therapy.

Even the blind complain about how terrible the UX of Blackboard is. Their opinion certainly holds more weight than whatever pathetic appeal to authority you're trying to make.

Pointing to some ink on paper isn't some magic spell that compels others to take you seriously. It only convinces people that already appeal to authority.

Accessibility is a whole other can of worms. I suspect blind people have issues with a lot of software.

And I'm not the one appealing to authority. That's the people saying the author is a CS professor so he's obviously right without actually engaging in the claims.

I brought that up because you automatically assumed it was just some random teacher that knows nothing about UX. Maybe if you took a second to realise it wasn't some random teacher, I wouldn't have mentioned it.

Engaging in claims? You immediately dismissed the opinion when you read about clicks. So don't try and say people aren't engaging the claims. You don't even use the software so how can you say he's wrong?

If the user is unhappy, then maybe take a look at the reasons. But you just dismiss someone else's opinion because it doesn't match yours without even seeing what he's talking about. And that's why everyone is bringing up the question of if _you_ used it. Because what the end user experiences is what UX focuses on. If you don't know how the workflow of blackboard, you can't say that 15 clicks isn't a nothing burger and someone who uses the software has an irrelevant opinion.

Describing his work as "its mostly non-user facing" isn't as valid as an argument as you think it is. CS professionals, professors included, tend to follow (not work in) multiple disciplines.

But by using your reasoning, does that mean any past critical comments you made on topics outside of your UX profession is moot? No, I think you mayknow stuff outside of UX. So, I don't understand your reasoning to immediately dismiss OP.

But back to UX, so user interaction and response has no meaning in UX? Isn't the X in UX for experience? How do you judge how useful UX is without feedback?

My understanding of UX as a web developer, is that if the user is unhappy about how long it takes to access something, then the issue should at least be looked at. Maybe it can't be improved, maybe it can. It's not something to dismiss.

Anyhow, have you ever used blackboard? It seems from other comments that you haven't. If not, then maybe your comments are the unnecessary ones? Because how can one say the UX is okay without ever using it? I have used blackboard and canvas for years. Their UX sucks.

I do agree that Blackboard is likely horrible. That's not in debate here. And their UX likely does suck. But knowing something sucks is not the same as knowing why that thing sucks.

And "number of clicks to perform action" is not a useful barometer on the usability of a product.

And I'm not dismissing his opinion because he's not a UX expert. And I'm not dismissing his opinion of the quality of Blackboard. I'm dismissing his opinion on why Blackboard is bad because he's focusing on the wrong thing.

Much like you and the other people trying to argue with me. They're trying to paint me as defending Blackboard. I'm not.

And I'm additionally saying that it's understandable that he doesn't know why it's bad because it's not his focus. There's no fault in that.

Whether or not I've used Blackboard is moot because I'm not discussing the UX of Blackboard, I'm discussing the use of the metric "number of clicks". It's a bad metric. Regardless of Blackboard's usability.

No, number of clicks isnt a _bad_ metric, but it isn't the most important. But when it comes to _UX_, user experience is important. If there are a lot of people complaining that it takes too many clicks, then it's quite possible that the workflow can be reduced somehow. And it's not just this one guy that has issues with Blackboard. There's been numerous end users complaining about usability for a long time. Regardless if how you feel about the number of clicks, it still is something to examine. Do some searches about it. Many, many websites will say clicks aren't the most important thing, but it is something to think about. Immediately dismissing someone because you don't feel it's important (even though other UX designers do) is a fault of yours, not the end user.

Reply on news.ycombinator.com