Ask Lemmy

I worked for a company that manufactured products and had been running on NetSuite as the ERP for about a decade when I got there. It had been customized and tweaked and worked pretty well for what we did over that time frame, with lots of time-saving automation. Shortly before I joined the company they were bought by a conglomerate and merged into a division with a couple other somewhat complimentary companies in the central US and west coast. Although it was supposed to be a merger of equals, it soon became apparent the west coast company had won the merger and was calling the shots. They were closer to our largest customer base and while our revenues were pretty similar, they shipped much smaller volumes and had much higher margins (they apparently at one point had a box that just had a Raspberry Pi or something similar inside that they sold for $5k/each). The big difference was we were a big name in a market with a lot of competition, so we had to be efficient and smart with our margins, while they were only big in markets where they had no competition; where they had competition they were often the last choice. While the plan was originally to move everyone to NetSuite, which already had options to run multiple companies/subsidiaries out of one instance, that was abruptly cancelled and we were told we would instead need to switch to Xtuple.

Xtuple was awful. NetSuite runs in a web browser but Xtuple opens multiple windows that look like something written in Java in the late ’90s. Want to copy some text? Unless it’s in an editable text field you can basically forget about it, and even if that field was once editable, many of them can never be edited again after the first time you save, or sometimes even as soon as you click out of it after your first time typing in it.

I don’t even know how much of it was Xtuple’s fault versus the company’s customizations. Where switching to NetSuite would’ve put all the companies in the same instance and allowed for one store to sell all the products, the plan to switch to Xtuple meant a separate server for each company, plus a fourth server to coordinate with each location. When you license Xtuple, you also get access to the source code and can make changes as needed. There was one guy at the west coast company who had total control of the software and no one else had access to it. It seemed like he used this opportunity to create the proverbial million lines of undocumented spaghetti code and guarantee job security. To try and help him with creating all the different Xtuple servers, they hired a consultant directly from Xtuple to create our instance, but when the spaghetti code guy came to integrate it with his part none of it worked, apparently because spaghetti code guy was doing all sorts of things in a non-standard way. This delayed our launch by 3 months because spaghetti code guy then went and rewrote the stuff the consultant did to make it work, and of course a lot of things still didn’t work right for several months after we launched.

Because Xtuple didn’t do everything NetSuite did, some functions were moved to outside software, like Customer/Technical Support to Zendesk. The built-in tools in NetSuite weren’t the best, but it was directly integrated so it pulled customer and product history in and could make an RMA directly in the ERP, so when the product arrived Receiving basically just had to push a button to check it in (the west coast company had never bothered putting their repairs in Xtuple much beyond listing the final price of the repair; they managed the actual repair process in a massive, unwieldy Google spreadsheet). Zendesk didn’t have that, so all of the data had to be manually entered twice, once in Zendesk, then again in Xtuple. We also didn’t put our old customer history into Xtuple, including even just a customer list, partly because the west coast company assured us they already had all the same customers as us (it turned out they maybe had a third of our customers, and that only counts business customers, not individuals). Since we had access to the code it seems like we should’ve been able to tie directly into Xtuple, but spaghetti code guy would only allow custom APIs he created. We never even got that to work because a year or so in the head honchos decided to move to Salesforce instead, so they spent a ton more money trying to make that work. When I left they still weren’t communicating, and people coming in from the conglomerate were starting to ask why millions of dollars had been spent on multiple transitions to rebuild functionality that still wasn’t working 3 years in. They were also cancelling the 4th Xtuple server to control the other 3 because they just couldn’t seem to make it work.

In the end there are a lot of things I don’t miss from that company, but I found Xtuple to be especially x-stupid. Still, I don’t know if it was the software itself or spaghetti code guy. Everyone acknowledged he was a problem when he wasn’t in the room, except maybe the CFO, but no one could do anything about it because it seemed like the businesses would completely halt without him.