What Can Be Learned From A Software Development Case Study?

It has been said that smart people can learn from the mistakes of others. Case studies are often created to compare one project to another. They are also used as single case examples of someone’s success. In this video, you’ll learn about how complex and detailed different software projects can be. From experts who truly understand their fields, learn about the struggles and intelligent motivations that can lead hardship into success. Watch this video to truly understand the challenge of making something appear to be simple.

Video Transcript:

Plexus software development case study – part 1

Ben (Alliance Software, CEO & Founder): Andy, why don’t you start by telling us who you are and your background and those sorts of things for this software development case study.

Andy(Plexus, Managing Director): My name is Andy Mallett. I’m the managing director of a business called Plexus. My quick background is I was originally in the financial services industry. I went into management consulting for a while and then, when I decided to leave that, I decided to start a business and wanted to do something a bit different that was a bit more challenging.

A friend and I had an idea that no one liked lawyers or law firms and perhaps there was a better way to run a law firm or a legal service business  than people had envisaged previously. So five and a half years ago we started Plexus with the simple idea that about 85c in the dollar that you spend with a law firm only goes to partnership profits or overheads. By inverting that paradigm and still having all the things you love from a law firm; top quality lawyers, but without a lot of the overheads and without a partnership model; we could deliver better value for money legal services than had ever been considered.

That model went very well and grew very rapidly. In the space of about three years we had about sixty of the top hundred companies in Australia using our services.

Then we came up with the idea, when we started to look at underlying legal tasks, we tried to figure out how to do the tasks better. The first part of our model is an arbitrage play around stripping out inefficiency, but actually we wanted to get underneath the task and say how we could do this task better.

We were applying Lean Six Sigma and process mapping methodologies to that and it became pretty clear that a lot of legal tasks were ultimately rules based. As you know, software is all about rules. We came up with what was, at the time, a unique idea; which was applying software to automate legal tasks. At the time everyone said it couldn’t be done. We spent two and a half years proving that.

Ben: Tell us a little bit about that. Tell us what you’ve built and what it does. Give us a bit of the story.

Andy: Going back about three years ago we started on this journey. My focus from my management consulting days is to focus on what is our core competency and then try to find other people who could bridge competencies as we move into adjacent markets.

Given our core competency was not around software development, we said we really need a partner to do this. There wasn’t anyone in the legal industry who had any expertise in this. There was one company in New York which was pretty early in their journey, but was probably at the time the global expert in legal expert systems. We went with them and they said, yes, we can build it.

The first use case was quite a niche application which was for running promotions. If you think about a consumer focused business, like a Qantas or a L’Oreal, win a ’round the world’ trip competition. I’m sure you guys have entered one previously. It’s got a lot of regulatory complexity about it because every state has different legislation that governs gambling laws.

So you need terms and conditions that comply with all those different laws and you need permits and provisions and a whole lot of other things. So it’s quite a mess. It takes two to three weeks to get the legal tasks done in the traditional format. A traditional law firm would charge you $2,000 for the service.

We had a client who said this is a nightmare for us. Could you come to us with a solution? We started working with this organisation in New York that had a bit of a platform. If you think about what WordPress is for websites, they were presenting their platform as a similar thing for building legal automation solutions on.

There were three things that happened.

  1. One is probably the individual on my team wasn’t capable of dealing with what was a complex software project.
  2. The second thing was their platform wasn’t as mature as they led us to believe.
  3. The third thing, as is so often the case with software development, the fringe use cases were immense.

So we probably spent two years, probably $400,000 or $500,000 on that journey, before we said we’re throwing good money after bad.

Question: I’ve seen that package. It’s a powerful package. What was the tipping point? When did you decide, ok, we need to move from a Software as a Service package to something we built ourselves?

Andy: We had an original business case around both productivity for us as the service provider, but also productivity for the end user. There were a couple of things.

It was clear that their platform was not stable. So we consistently had issues of falling over and bugs; issues that shouldn’t be present in the platform and that’s when we thought of building bespoke software. We got frustrated with that relationship and they weren’t living up to what they had committed to us, which was problematic.

The second thing, if you think about it, because of the fringe use cases our lawyers spent a lot of time redrafting the output to make sure it was fit for purpose. If you can imagine a scenario, it may be there are a hundred different prizes, different draw dates, there are six different steps to enter. That sounds all very easy to put on a flyer or an ad. But when you think about the layers of complexity of delivering that outcome, but also ensuring that all interfaces agree with what is a couple of hundred regulations across Australia, it’s very difficult

Ben: I know the project. What they’ve achieved; they’ve created a platform that non lawyers and I mean admin people in these organisations, can use to effectively produce incredibly complex documents that cover a whole range of areas. The complexity has been months and months in analysis really.

You’ve decided I’m going to custom build this thing. What is your thought process for this software development case study? Tell us about it. What were you thinking, what were you worried about? Give us a sense of your world at that point.

Andy: It’s funny now. We had our quarterly board meeting last week and one of the guys at the meeting said, this is an amazing business. I said, Pete, eighteen months ago you told me to shut this down. He clearly did not remember saying that.

As managing director I was copping a fair bit of heat because I had spent a lot of capital at that stage and we had made commitments to clients also, so it was not a situation any business leader wants to be in. Then doubling down and saying we’re going to add another layer of complexity, that we’re going to do a top down build on this, was quite concerning.

I said to the board last week that the only reason we got here is through tenacity, but also I was spending all that money. I didn’t want to be beaten. I was thinking if we spend the same money again and still don’t get to the right outcome it’s a really bad problem and not a good story. When I was a management consultant I spent a lot of time advising CFOs. I remembered that, I think our data point was, 72% of IT projects never realise their business case. They’re internally focused projects, you don’t have product market fit when you’re selling software to worry about.

I was deeply concerned we would repeat history there. What do they say, history doesn’t repeat itself but it does rhyme. I was once bitten twice shy, so I fortunately had a CFO at the time, she was an IT consultant and one of the best professionals I have ever worked with. I asked her to lead the project. She is a highly systematic person as you can probably guess, being both a CFO and IT consultant. We did a pretty detailed analysis of the risks and the options. We had a really good product manager on board by that stage.

Going through the process we said the only option if we are going to do this again is to build custom software to do it. We did three months of analysis of do we think it can be built. We did an initial project with you guys (Alliance Software) to validate our hypothesis from a technical standpoint. The answer came back, yes, it can conceivably be done, it’s just not going to be easy.

Ben: Interestingly, the challenge here wasn’t what I would call a purely technical challenge –  we knew we could take data and create documents, that was never in question. The challenge was could we do it in such a way that it would make sense? That someone would be able to look at it and be able to follow the process through with what are mathematically speaking millions and millions of permutations and combinations that the thing can go through. When we talked about validation, that was really the piece we were doing. We made something like 450 slides of wire frames by the end of it. We pushed it pretty hard.

Andy: I’m not famous in our organisation for being what other people think is reasonable; I think I’m entirely reasonable.  I said this has to be easy enough that someone with no experience of promotions can do it, yet complex enough that it will address anything within 90% of use cases. Our rough estimate is 20 to 25 million different outcomes that it can generate. It’s not straightforward in terms of that.

Sometimes it isn’t enough to have a great idea. Even the most experienced software developers go through times where the logical appears almost impossible. We can learn so much from the challenges that other developers face. Sometimes, the answer is to press ahead and realise that making something simple and logical for the user means something incredible challenging from the programming end. By running a test of processes to prove if an idea is feasible, we may find that it is in fact worth living up to the project challenge.

Do you have a great software idea that you want to pursue, but need input from experts to prove whether it is better left as a dream, or can be rendered into reality? We’d love to hear from you, so contact us right away.