Why Business Automation Projects Fail — and How To Avoid It
This article originally appeared on the Next Matter blog here.
Large-scale automation projects fail at an appalling rate.
Legacy system upgrades, new system implementations and rollouts, multi-departmental system integrations — these are just a few examples of enterprise “automation” initiatives traditionally associated with big resource consumption and implementation timelines spanning months or even years.
Many management teams still accept this routine as “a necessary evil of doing business.”
Yet the research shows (now more convincingly than ever) that projects scoped and implemented along these lines are prone to deliver sub-par results — by some estimates, over 70% of the time.
And that’s not just a financial failure. Most of us have either participated in or witnessed a major automation initiative going down the wrong path. It’s a wildly demoralizing and time-consuming experience for employees and “task force” teams, who walk away from projects not sure what — if anything — was achieved.
Realizing that businesses — particularly those in growth- or scale-stage — can no longer afford such costly failures, let’s look at the former “best practices” you’ll want to avoid if you hope to see success and return on your automation investments.
How operations organizations screw up automation (DON’T do this)
The first problem is, the traditional “waterfall” approach to automation project management isn’t primarily concerned with achieving results — at least, not in the beginning.
Typically, project teams are sent running in different directions: compiling requirements, investigating new technologies, drafting up solution proposals, and only coming together for the “necessary” project team update meetings. The idea, of course, is to assemble all of the research, recommendations, and proposals into a working blueprint for implementing a solution — and only following the implementation of said solution, seeing some sort of performance results in return.
This is the wrong approach for two reasons:
Firstly, the traditional intent to “capture and plan for every angle” from the onset is inherently flawed. No project manager can possibly foresee every single consideration, let alone plan for them. Without a more iterative approach, this can cause minor hiccups in the best case scenario, and completely derail or kill a project in the worst.
Secondly, by putting emphasis on planning and implementing the “perfect” solution, you run the risk of dragging out the timeline so far, that by the time the desired solution is finally implemented, the needs and requirements of the business are likely to have changed, and the team — not having seen any impact from their project work — have moved on to focus on other priorities.
Let’s use, for example, the classic CRM upgrade.
Suppose your goal is to double monthly quotes sent by automating the sales operations process. The proposed solution is to upgrade your company to a CRM with greater automation functionality.
With a traditional approach, you might have one team research and test software, another team analyze the current sales workflows and technologies used (Email, chatbot, phone, another develop documentation, and so on. But only during rollout — already 3–4 months down the line — do you learn that your reps actually despise using the new CRM, find their own workflows much more efficient, and refuse to make the transition.
Sales management supports the reps, the management team goes into gridlock, and just like that, the project fizzles out — months wasted of multiple employees’ time, and not a single result achieved.
This exact scenario has derailed CRM rollouts at Fortune 500 companies, all the way down to family-owned retail businesses. But it’s just one of many possible failure scenarios when businesses take a traditional approach to automation.
So, how can you avoid throwing your next automation project in the trash bin?
How to implement automation the right way (DO these 4 things)
Regardless of the exact business function, your approach to automation should never begin within the framework of legacy system upgrades. It is precisely this “in-the-box” approach that is usually associated with huge costs, long timelines, and sub-par (or completely missing) results.
Instead, take one big step back, and take the following 4 steps, to implement reliable automation and see real results, fast:
Step 1: Create an Automation “Mock-up”
Start by considering the ideal automated operations processes you’d like to see, and what that might look like. Don’t think about systems or requirements at this stage.
Simply ask yourself:
“In an ideal world — technology constraints aside — how could this operations process work better?”
And then, the follow-up question:
“Which parts of that process should be automated right now, which would already make an impact on performance, reliability, simplicity, etc?”
Document your thoughts and your “ideal” process as plainly or as detailed as you like, in whatever format is easiest for you. Pen and paper work just fine. You can also use the steps and template provided in this free automation playbook.
Step 2: Find a No-Code Automation Tool
With your “ideal” process mock-up in hand, look into a no-code automation tool, that you can either set up quickly on a free trial account, or purchase on an entry-level license. No need to bother with assigning a task force to do the research — all the documentation you need is on that piece of paper in your hands.
No-code automation allows you to bypass the entirety of IT project management, and doesn’t require business users to have any additional coding or development training.
That means that you — the operations leader — can implement the automation all by yourself, without having to speak with any IT consultant, or engaging with a “task force” internally. It also means that instead of focusing on system compatibility and requirements, you can actually prioritize automating things that will drive immediate results
Be mindful, however, that not all process automation tools are made equal. Make sure to look for these features and functionalities in your process tool, at a bare minimum.
At this time, you should also nominate your process MVP “pilot” team and inform them that you would like to implement an automated process, and would like their support in testing and providing feedback.
Step 3: Build Your Automation MVP
Build your first process MVP following the “pen and paper” (or however documented) framework you created earlier. Workflow automation tools are extremely powerful, but you should resist the temptation to implement every little functionality and detail from the start.
Instead, focus on building out your process MVP end-to-end, and automate the basic steps. Make sure to cover things like:
- Create an automated step for each step of the process (“what to do”)
- Provide work instructions, checklists, and necessary data references (“how to do it”)
- Assign team members and stakeholders to process steps (“who does what and when”)
Just make sure that your process is fully-built, from start to finish, so that you can already run real-world tests on it. This should take no longer than 1 business day to complete.
Note: integration of tools and systems can also be added in your MVP, but isn’t necessary to start testing.
Step 4: Run your automated process, iterate, and improve
Start running your automated process alongside your designated pilot team, and observe what happens. In part II of the operations playbook, you’ll also find an easyframework for evaluating your automation pilot.
Odds are, if your operations process was previously coordination-heavy (i.e., requiring lots of email, phone calls, chat, etc), then the automated version will already show increased reliability and efficiency. These gains should be realized by your pilot team almost immediately.
As you continue running instances of your new, automated process, you and your team will also begin to spot opportunities for improvement, and further gains. For example:
- An additional meeting or coordination step, which can be eliminated entirely
- A reference to data to or from another system, which can be sent or pulled automatically
- An opportunity to provide deeper context in work instructions, yielding more reliable work results.
Above all, the process continues to run, iterate and improve. “Quick win” results were realized from nearly the onset, which — with time — can evolve into game-changing results.
And if something isn’t working right? You just fix it yourself, or you contact your no-code provider, and they can help you directly. No need to involve IT, consultants, or a project task force.
Making your Automation Project a Success — it Starts With Common Sense
Almost every business leader has sat in an IT project “task force” meeting and thought: this doesn’t make any sense.
That’s because the traditional model of IT project management — at least, when it comes to operations automation — really doesn’t make sense. Especially when, for a fraction of the time and cost, you can implement a real automation solution faster, and one that you can build and modify yourself, without involving IT,
consultants, or a task force to work on it.
Indeed, by focusing on getting results fast, you place the emphasis on the pure intent of automation itself — that is, doing things easier, faster, or more reliably by automating components of a process or task. The automation speaks for itself, rather than the other way around, where a room full of individuals talk about what automation will — eventually — do for them.
Attempting to build complex automation in fast-growing business environments is a humbling experience. Many, many operations managers know this challenge to be an everyday reality. And very few leaders or businesses have figured out how to do right, consistently.
Regardless of your track record on automation technology rollouts, you can take an important first step in the right direction by forgetting the traditional approach to IT project management. Managers believe they will be able to plan for and control every circumstance months in advance, but they can’t. And they’re wasting both their time and their employee’s time trying.
Using a results-driven, no-code approach, however, managers can implement a cycle of discovery, improvement, learning, and iteration, enabling their teams to produce actual results fast, and feel an even deeper connection with their work. And best of all, it ensures that no one’s work ever goes to waste again.