Automation over Bells and Whistles
Sometimes it is hard to see the forest for the trees...
Ask yourselves: would you rather have a new drop-down menu option, or just get the task done automatically without having to log in to a UI and choose a menu option at all? Automation is the new Usability. More automation services startups to reduce business-diluting tedious tasks, please!
As a former startup Engineer and Product Manager I know the joy of building something new, and how great it feels when you, together with the engineering team, are working on something that will beat the competition and/or enable users to do a new thing they haven’t even imagined they could do.
A decade ago, a helpful new feature often referred to building out something new in an end-user friendly UI, or a new option for an administrator to choose. For the last 5 years, product management trends have been about persona-based product design. If you are new to the challenges of persona-based product design for product organizations and products that are serving more than one persona, let me provide a short example:
Imagine a product that needs to serve 6 different personas (security admin, ops admin, application developer, business user, data scientist, governance auditor, etc.). Imagine then, the feature flow into this product and the contention of resources in engineering to get all these features done - including quality assurance and documentation of course. In reality, often none of the personas ends up really happy because:
They only get parts of their workflow supported in a release, and not the full end-to-end workflow due to competing priorities and lack of resources.
Optimizing user experience is hard, takes many people involved (Designers, Front end, Back end) and it rarely ends up with an intuitive enough result across these responsibilities.
Different companies have different processes and definitions of personas/roles that need to be served differently through the tool, so the “persona” often gets ambiguous
Either way, I think most of you would agree that persona-based product design is hard, ambiguous, and resource hijacking for multi-persona-based products.
As a startup product owner, you know adoption is key and usability is key to adoption. But what many startup product owners miss is that usability does not necessarily have to be served through CLIs or UIs. Often, way too many resources are spent there when usability actually could mean a whole different thing!
Sometimes it is hard to see the forest for the trees.
Time to value. This hunger for faster delivery, more business agility, quicker releases, quicker fixes, agile movement, continuous delivery, micro-services, cloud-first - they are all movements to a faster pace of product features into production, which means more value for customers along with faster and faster innovation than competition.
If not done right, this comes with risks for:
poor product quality
costly resource utilization
overworked, stretched teams
And probably risks in other areas too.
What really would be valuable (hence useful) to any organization following these trends would be to reduce the pains of human involvement at all. Usability could mean relieving overloaded humans of making difficult decisions in stressful times. Usability reshaped as automation would help prevent mistakes caused by human error, often brought on by fatigue, and would help us all get out of the option-hysteria UI design. Why not take a different perspective on the whole usability and persona-based design topic? For any and all parts of the time-to-value task chain strive to not having any human involvement unless absolutely necessary? Serve the persona’s workflow by not having them to do the task at all!
Tedious repetitive tasks should be outsourced to services. Services that require minimal human interaction.
Carefully evaluate:
A human decision of pushing code through or having a system evaluate if it's ready?
A human having to check if all security scans have been made, or having a service ping you if they didn't run for each code segment?
Having a service auto-detect and assign actions when things go wrong, over waiting for the next available first responder to do so?
Having a system updating lagging documentation, over waiting for the first customer ticket.
and so on...
We are in an era where pattern recognition is a viable production feature of any ML-driven SaaS service, so by all means, it is time to outsource to services; it is time to automate.
Here is my ask to all UX/Design/UI eng teams out there: think automation first next time before putting another UI addition or CLI command in the product. Automation is the next era of happy-customer defined usability. As a Senior Systems Engineer connection of mine recently put it:
“I’d pick automation over new bells and whistles any day.”
Let’s focus on automation as far as possible, over expecting a human user to be involved. Let’s not put the burden on someone to have to see something, to click a button to act on it, and to morph into a human bottleneck to move a process along. I believe this would lead to healthier software, healthier teams, and healthier customer interactions long term. Automation is much more helpful than providing a user with another option to choose from. Automation is not a replacement of human resources; it is a relief to allow humans to focus on the most business-valuable tasks!