Enterprise Software and the SaaS Contraction

Thinking about both recent changes in the technology landscape and the state of enterprise software solutions, I suspect we may be about to encounter some major industry changes in the next few years. I believe these changes will see a contraction of SaaS offerings, enterprises simplifying their business software and systems on their modernisation journeys, and leaning more into a streamlined approach.

What do you mean? 

Well, firstly, let's examine the state of enterprise software solutions. 

There are several key domains and suppliers: 

• Enterprise Resource Planning (ERP) Software (SAP, Oracle, Microsoft Dynamics 365, Infor, Epicor) 
  

• Supply Chain Management (SCM) Software (SAP SCM, Oracle SCM Cloud, Blue Yonder (JDA), Infor SCM, Manhattan Associates) 
  

• Collaboration Software (Microsoft Teams, Slack, Zoom, Google Workspace, Asana) 
  

• Customer Relationship Management (CRM) Software (Salesforce, Microsoft Dynamics 365 CRM, HubSpot CRM, Zoho CRM, SAP CRM) 
  

• Human Resource Management (HRM) Software (Workday, ADP, SAP SuccessFactors, Oracle HCM Cloud, BambooHR) 
  

• Business Intelligence (BI) Software (Tableau, Power BI, Qlik, Looker, SAS Analytics) 
  

• Content Management (CMS) Software (WordPress, Drupal, Sitecore, Joomla, Adobe Experience Manager) 
  

• Project Management Software (Monday.com, Trello, Asana, Jira, Microsoft Project) 

On top of this, over the last 5-10 years, to deal with the volume of SaaS vendors popping up in the enterprise IT and software domains, the need for middleware to connect all of these disparate SaaS systems has been huge. This has enabled a new segment called API middleware, which acts as an intermediary between systems, including those providing hosted data storage solutions. The top vendors in this space include Kong, MuleSoft (Anypoint Platform), and Apigee (Google Cloud), offering solutions that manage, secure, and streamline API integrations. This has led to the term API going from a niche IT term to common enterprise lingo. 

Enterprise SaaS offerings have not actually replaced the core legacy systems in the enterprise. This means there is still a large amount of technical debt, and on top of that, organisations are burdened by the complexity of maintaining both legacy systems and new SaaS applications. For instance, a study found that 92% of organisations are struggling with technical debt caused by legacy systems. 

In the current environment of mixed legacy and SaaS systems, employees are frustrated by switching between multiple platforms. The integration of SSO (Single Sign-On) and Entra ID has helped reduce this frustration by allowing users to log in once to access several applications, including collaboration tools and file sharing platforms. Entra ID (formerly Azure Active Directory) streamlines secure sign-ins and user management, improving both access and security. 

Now, let's take a look at what's happening in the technology landscape... 

As we all know, enterprise software development companies have entered the "There's an AI for that" phase of the hype cycle, which evolved from apps to SaaS. Now, everyone is hitting the "generate" button on the next huge AI, but there are really only a handful of players with significantly high-quality models. The fact that Apple had to 'partner' with OpenAI was a jaw-dropping moment for me. 

Anyway, if users are already frustrated by the volume of systems, do we really want all of these systems and extra Gen AI systems to "empower us"...? 

Add to this, the flip side of the coin: it's never been more important for any organisation to increase their cyber posture to mitigate the threat of attack. When it comes down to how you actually do this, the answer is pretty simple: better training and governance. However large your threat surface is, the higher your risk and the greater the cost to train and govern. So, what should CIOs be doing? They should focus on cutting unnecessary business software solutions and simplifying operations. The fewer blood-sucking SaaS subscriptions, the better. 

So, there are AIs for that, there are too many systems, and we're really still copy-pasting Web UIs into the mainframe. So, what should we do? How do Australian enterprises cut the SaaS fat, leverage AI for optimisation, and still take on the world? 

Well, do I have the answer for you? No, I don't... but I can share what we've done as an organisation. One of our core values is balance with direction. We've empowered specific team members to maintain their systems, given them a set of core tools to use, and made them responsible for the governance of these systems. 

Here's what we did: 

• Set out in our organisational manifesto that we have 2 core systems to empower our team: 
  

1. GitLab (free - open source) - We use this for: 
   1. Project Docs (internal and external) and store all knowledge in Markdown (.md), which allows users to write and share documentation, making it easy to create well-structured and readable text with minimal formatting effort. 
   2. Project boards: Action, Dev, Compliance, Incidents, and compliance. 
   3. Continuous Integration and Development Pipelines (CI/CD): automates the process of building, testing, and deploying code (.md docs, files, and knowledge), enabling continuous integration and continuous delivery in software development.

1. Codebots (our own model-based platform engineering toolset) - We use this for: 
   
   1. Creating models and meta-models of customers and our own systems to collaborate on. 
   2. Reusing the templates associated with models to drive momentum (code and projects). 
   3. Pipelines: Linking multiple models up to GitLab and empowering teams. 
   4. This gives us the ability to model any system in any technology.

• Set up our Information Security Management System (ISMS) to 27001 standards and assign accountability and responsibility within the system. 
  

1. This means, if I want a system, I have to own and audit that system. 
   
   1. Individual teams pay for their own technical debt by governing their cloud services and SaaS. 
   2. This enables us to maintain ISO27001. 
   3. Personally, in doing so, I cut about 50% of the systems my team was using. 
   4. As an organisation, our SaaS bill went from about $150K per year to <$25K (that's a whole new role in the organisation for an SME). 

That's where we're up to... What else would I love to do? 

Well, as part of the ISMS, we developed our own responsible use of AI policy. The way I see the market is that the open-source versions (Facebook LLaMA, etc.) are maybe 12-24 months behind the industry leaders. The price to give everyone in the organisation access to OpenAI is too expensive for an SME. We need to balance how we use it, and I would love to link it up to our GitRepos to further empower the team (just need to make sure this doesn't make us dumber but is only a time-saver). 

Therefore, I would love to install and give the team access to our own internal LLM. 

How does this compare to the rest of the market and SaaS contraction? 

• The SaaS contraction will reduce governance overhead and risk. 
  

• Focus resources back on your core system. 
  

• Focus on integrating new technologies into your core systems, not adding new systems. 

Less is more!