Pros
Lets get the good stuffs out of the way. The ceo seems capable, ambitious and wanting to build a profitable software business (instead of bubble it up and exit) in an old industry (hr, payroll, superannuation and eventually the goal was to become a neo bank). Some backgrounds of this business. Flare started our as a SaaS hr software business, it is free to use (the hr part) and the revenue came from selling employee benefits (e.g. movie tickets, vouchers etc). Later on the company partnered with westpac (majority of funding came from westpac anyway) and started selling pre packaged BT (part of westpac) super products and then various other super products from other providers. There were talks about going into pay day lending, bank accounts and eventually becoming a bank, but that seemed too big at the time and perhaps the intention was to draw bigger investment. The team was highly capable of selling unfinished products, for a “startup” this is important and was executed extremely well. However I believe they did a too good job and the delivery failed epically (more on this later). There was another stream of revenue which proved to be not so much a good idea. The company partnered with a sap provider to be able to get into payroll quickly. However 12 months later, only one client went live with urgent production bugs needed to be fixed on a daily basis. Also Flare charged $8 per employee form its customers and paid about $6 to its sap partner.
Cons
While the business side seemed ok, the dev side was as bad as it gets. First, there was huge power struggle with in the team, no one listened to anyone else, and all the decision making were coming from the top. Even though the culture was supposed to be that the decision should come from the team. There was close to zero collaboration, if one person came up with the idea or wrote the code for this area, nobody else can touch it nor willing to fix bugs. There was a lot of blames even though the culture was supposed to be blameless. There was also a lot of sentimental thing going on within the team for no apparent reasons. So on the surface the team seemed cohesive but in fact everyone disliked each other / and their code. The software was not written from scratch, it was a pre packaged / pivoted from previous venture and hacked into a hr solution. This software were written 15 years ago, using all the possible out dated technologies. If code generation sounds familiar or xml configuration, you got them right. Everything was code generated. From database tables to database objects, and domain objects, then controller dto. How about javascript objects that’s used in the front end, no problem at all. Want to add a new table? No problem, just copy paste and modify my T4 template, viola, all generated for you. The team seemed to be extremely familiar with xml and xlst. XML configs were saved into database, html were generated from xml and xlst. There was no concurrency control, no proper business rules (bunch of nested if / switch statement). Ever heard of domain driven design? No. What about UI composition. What's that? How about microservices? The terms "startup" seemed to be used frequently to defend bad software / design flaws / bugs. The team seemed to have no prior accounting background in designing accounting / payroll / hr software, just drop the database field onto an already massive web forms (how about a web page containing 100+ text boxes, dropdown menus etc), pick any available space on the screen, drop it there, generate the backend code and be done with it. There’s no concurrency control in code, meaning that if two customers editing the same employee details, the last save would overwrite (with old data) the previous save. And imagine you can not work out how much annual leave you have in order to decided whether you can take the next holiday. The testing got most of the blames for not able to report bugs. Devs don't do testing, they passed it on to the tester and blame the tester for bugs. Bad software affected implementation / delivery. Implementation and delivery took on average 3-6 months. The staff turn over was quite high, there were a lot of hires and fires within short period of times.
