The scapegoats guide to organizational 'transformation' [DRAFT]

Digital transformation. Evolving into a data driven organization. Becoming an AI-first company.

With every technological wave a bevy of consultants (like me!) line up to “transform” organizations.

From Accenture

Accenture is their reinvention partner of choice

From McKinsey

Digital transformation is the rewiring of an organization

Leaders recognize their problems and seek outside help.

Hiring an outside expert feels like accomplishing something. But its an input, not an outcome. Problem delegated, it might mean the org (or leader) forestalls their own personal growth needed for the “rewiring” to succeed. It’s tempting for leaders to go back to their busy, overwhelming schedules, and assume all is in hand.

You, dear reader, might suddenly find you’re this alleged transformational agent. Either as a leader or part of a team. The company pins their hopes on you.

Sadly, all too often, after hopes run high, the expensive transformation backfires. Now instead of one of the cool kids doing the new, hot thing, you’ve become the scapegoat that caused the mess.

Could you have done more?

Maybe this post about my reflections on my years “transforming” organizations might help you. I’ve been around the block and been burned.

Idealization and devaluation cycles

Any new hire or project goes through a “honeymoon phase” - we see only the promise. We rarely see what can go wrong.

But when a new transformation (person/team) goes through a much more aggressive version of this: idealization followed by devaluation.

Idealization:

Idealization refers to a person’s tendency to assign exaggerated positive qualities to a person, place, or concept

Of course, during the idealization phase, we put on all the hopes and dreams of our company’s “AI transformation” on one person, team, or effort.

Frequently, this doesn’t work out. Nobody can meet such lofty expectations. So inevitably comes the devaluation:

…devaluation refers to the act of assigning exaggerated negative qualities while disregarding the good. During devaluation, flaws, weaknesses, and negative traits take center stage, and positive qualities are completely ignored.

You were once the shining knight coming to save the day. Now you’re the scapegoat we can throw under the bus. And a new idealized transformation is coming online (likely to be devalued in the future). Buck successfully passed.

Idealization: transforming the ultimate ego mindf*ck into a tool

Idealization can push you into the lazy trap of believing your own hype.

I have gone though this cycle a few times, only to look back with regret at how my ego was seduced into believing that I, alone, had all the answers.

My failure goes beyond “believing my hype”. In reality, I fail because I believe that the work needed is my work, my chance to shine, when it’s their work. You need to lead with tough love, and push leaders to do their work for any transformation to stick. During the idealization phase, you have maximum capital to ask for things.

• Be very clear on how often the transformation you’re attempting fails. Often.
• Be clear it will fail if you don’t see X from their side (leader needs to arrive at X decision by Y date, etc)
• Be mentally prepared to leave early if you don’t see the change you need (prepared internally, don’t throw around you’re going to leave)

You’re countering the delegation trap, that it’s all your work when the leader is the one that needs to change.

This coaching starts from the very first interview or sales call with the client. You have a short window (maybe two months?) from this point until your idealization runway is cut short. You have their attention for a short while and you need to be prepared to push as far as you can.

Humility can be a trap

Another trap is humility. If you’re also like me, you often assume you DONT know what you’re talking about. Maybe you shouldn’t push to hard on X. Maybe how they do things IS the right way.

Organizations put so much pressure on team members to conform. You might feel you need to fit into their mold like a good little soldier.

This one is hard - humility is one of my core values. But I have to know I’m not paid to fit their mold. That’s not transformation. With transformation, they want your opinions, leadership, and perspectives. Strong opinions, held weakly.

If you do end up just conforming to be a “regular employee” in their organization, it won’t take long to see you’re way overpaid for that kind of role.

When a leader cannot tolerate someone not fitting the mold, it’s a red flag. You have an expectations mismatch. You’ve stopped being a thought-partner to that leader, and become an employee.

Show don’t tell (Ship don’t yip)

You can’t just be an ideas person, you have to be involved in the hands-on, nitty-gritty work of the organization.

A transformation fails because it sounds big, abstract, and hard. With hands-on work, you make success concret. You demonstrate what success looks like. This reinforces all the hard, inner work you’re asking leadership to do by connecting their tough decisions to tangible results.

How do you do this?

Your instinct might be to lead the team through a gradual, waterfall implementation of each supporting bit of “infrastructure”. Only after the team has all the right tools, and all the skills, should you deliver to customers.

For example, to ship a state-of-the-art recommender system, you feel you need to teach the team everything about recommendation systems. You need to build an A/B testing tool. You need to gather telemetry on user behavior. You need to staff up a team to handled all these parts.

All before a customer sees any value.

This invariably fails. Remember you have a short runway (see idealization above). A few months of being the golden child. Go out to 9 months to a year, the entire market has shifted with new priorities taking center stage. Leaders who hired you moved on. Your carefully crafted, ideal version tools don’t matter because now we do LLM as a Judge etc.

A better way is to ship improvements from day one. Using crappy, flawed tools and processes, building on the team’s existing, limited knowledge. SIn search, recommendations, etc this means doing grug-brained evals. Hacky, broken ways of testing that maybe improve things 5% towards an ideal. That big up front project and training? Almost certaintly YAGNI (you-aint-gonna-need-it).

This works because

(a) you’re shipping

(b) you’re building on team’s existing knowledge + practices

(c) the team will learn more from concrete implementations shipped to prod

(d) the platonic ideal of your problem (a platonic ideal search, recommendation, AI system, etc) has rarely existed in my experience. What exists are concrete, specific, implementations.

Finally you, the organization’s newb and golden child, have leeway to make mistakes others do not. If the organization is stuck in analysis paralysis, you at least can plead ignorance, use your capital, and try shipping a few things.

OK but then you’re just letting the organization depend on busted, old ways of working? It’s letting them cheat and not actually do a transformation?

Don’t worry, you hit diminishing returns unless you can incrementally improve infrastructure as you increase your solution’s complexity.

Growing tooling to do work and the work itself go best hand-in-hand. Teams get stuck because they think they need to be 100% better in one fell swoop. When in reality, they need to get 5% better with every feature shipped.

Building allies

When getting into shipping mode, find kindrid spirits in the orginization. Employees eager to learn about the new, hot thing. Odds are they are just as skilled as you, they just don’t have the permission structure you have. You can unleash them.

Letting them participate and take the credit will help you build momentum. It will solidify the foundation of a bottom-up transformation, making it easier to ship. It’s also a great feeling to let others succeed and share the acolades.

As you do this, you’ll discover how people currently ship. An existing path to production exists. Use that instead of being a special snowflake. (Don’t build a lot of new infrastructure, etc).

Getting leaders engaged

Your ability to influence the organization is capped at the highest level leader participating.

When you ship, you can get leaders out of delegation mode, and into participation mode. Transformaiton doesn’t happen unless they’re leading. To do that, bring them into the fold of the changes you’re spearheading. Let it become their way of working, not your way of working. Take your ego out of it.

Encourage them to steer. You built a dumb way to building recommender systems. But maybe there is another alternate path that they see? After all, they have the organizational expertise. The concrete implementation for this company is far more important than abstract “best practices”.

Unlearning

Much of what I’ve described so far depends on applying jedi-mind tricks to yourself

• Go against your ego by not making it your problems, make them their problems. Your task is for them to get out of “delegation” mode and instead for them to learn and change.
• Use your permission as an idealized golden child to empower others with skills similar (or exceeding!) yours
• Go against your expertise by being OK evolving from simple ways to gradually more complex ways

The application of expertise to “transform” requires setting aside your desire to own everything. It requires you to unlearn everyday:

In pursuit of knowledge, every day something is learned. In the practice of the Tao, every day something is unlearned

• Tao Te Ching, Ch 48

You exert influence, not control

While you can push leaders to do their work (and help teams evolve) you don’t have complete control over the outcomes. In the end, you can still fall down into some form of devaluation. I’d even say, to some extent, it’s inevitable. People leave. Markets change.

If you’re successful, the transformation you instigated might have success. If it’s their transformation, and not about you, then eventually they will congratulate themselves. Your role may get omitted from that historical account. Accept that’s likely.

Additionally, it may not be the exact transformation you/they initially imagined. They may be a few steps more mature in AI / etc. You might have positively changed a few people’s careers. Your inner expert might be offended by solutions they landed on.

Learn to accept and move on.

Your job is to exert influence, not control. What you’ve helped create is the legacy code needing to be “transformed”, part of a conversation. Don’t get married to what you shipped (the devaluation will only slip deeper). May you live long enough to become the villian.

Accept that you’ve had your impact and another wave of transformation may come. Ride off into the sunset with your head held high. You control your own narrative, and they need to control theirs.