More thoughts on System Thinking in Product Management
Find the way to notice interactions between pieces and don't let everyone else down
Links:
The final quarterly meeting is in full swing.
We have been forced to come to another city to stare at the projector board with some long-ass .pptx prepared by ourselves. As everyone’s seen it before (we have sent our slides in advance), there is nothing new here.
Another meeting costing us all time in terms of commute, opportunity cost and the hotel with all the premises to host the whole thing, another 30k down the drain to achieve what?..
For some of my colleagues, it is a getaway from the ‘normal’; they are excited to escape their hells at home, the shouting family members, 4-8 walls (those with smaller income streams who can’t afford to have more than 8). They are laughing. They are discussing the new game one played while another group is debating a new model of VW they want to buy. On credit, of course; who doesn’t need another credit, guys?
I am trying to hide my lousy mood by looking at the gathering. This is where my might-be bonus disappears, with every sip of that ‘free’ coffee produced by an ancient rusty machine in the conference room (swiss quality in its glory).
They can’t wait to start drinking afterwards, and some are passionately licking their lips in excitement.
I am focusing back on the speaker, who’s been another PO for a while now. For a while, because I can be considered a freshman in the company, he’s been here for some 6-7 years. Unthinkable.
He raises a list of issues they were able to fix this past quarter - a revenue stream from X on our page. Wait a moment… It is my scope too!
I have been there a lot, even before I got into Product Management and Scrum.
We tend to imagine our workspaces as a beehive full of smart virtuosos focused on one thing - what’s the target at that particular moment. Dedicated, calculated and eagle-like focused.
But if we look under the hood, we see a bunch of folks with different agendas, skills and… well, pay.
Given the variety and these factors, we tend to build up our domains, our silos, our safe heavens optimized for that one thing, the way we see fit.
Some folks are running ads, and some do fresh stuff so that Bing pushes them to the top, and others, well, others give their shot at squeezing every transaction through upsells, cross-sells, “you definitely need that hoodie”-sells.
Some do it through margins, and some do it through hidden fees.
And the most proficient Product People (capital case) are good at merging it all together.
The higher the level, the better the symbiosis.
But I wouldn’t write this post if it worked well.
Those pieces above work well in separation, but the interlinks, those complex interactions between the elements of our systems, are pretty often disjoined and heavily skewed.
Moreover, as we cannot see the whole, we race towards local optimizations, which dumps our problems onto other people’s domains.
Say, we want to increase the revenue of our Y module located on a homepage (team 1).
Being reactive thinkers, we see the problem - it doesn’t earn well, so we do our best to optimize it based on our locally available data - for that particular functionality.
And, quite often, we succeed.
The revenue goes up, or so we think, but what we do is the transfer of our problems to other folks.
Say, the new version of that product made by team 1 weighs 5 megabytes more, which makes the page heavier.
The danger is slightly apparent to us, but we roll with it. The marketing told us so.
Someone will have to pay for that, in this case - that might be the team responsible for CRs on our mobile instance or… the guys doing their best to make us searchable through Google.
Search engines take their time to index you, and the effect of that change may take months.
The scoring is screwed, and here goes the penalty in terms of worsened Core Vitals.
Weaker Core Vitals mean our ranking slowly but surely goes down (note: I am well aware it ain’t that simple; it is still a theoretical scenario, nothing more).
Lower ranking means decreased exposure, which leads, in turn… to fewer clicks bringing us conversions.
The system reacted with a significant delay, but what our team 1 did there was a simple trick of cards; the problem got relayed to some other unhappy nerd behind the laptop. That’s his job now, for the next quarter, that is.
Another example would be enabling reviews on your eCommerce website.
Trivial.
Maybe, if you are starting out.
If you happen to have it running for a while now, you might encounter evident and not-so-obvious interactions between the elements.
Your content managers have their targets, so naturally, they tend to push some products here and there through promos and ads.
Upon release of the feedback functionality, you might encounter one of the scenarios:
Some ‘pushed’ products won’t sell anymore. The actual experience might be subpar (the product itself or your service; for some people, there is no difference between those; it is all about the experience). This will migrate your traffic problem (review entries were used for Google Ads before to optimize the CRs) to someone handling sales of certain products; some of the positions in the catalogue might go down.
If there is no incentive to leave ‘good’ reviews, you will end up having a skewed distribution, with the majority of reviews describing your bad quality and lousy service. Expect lots of moaning and sweating here.
If you are in the marketplace business, your users might misunderstand the fact you are only a middleman and rate the transaction based on the combined perceived value inferred from the product quality & your seller’s service level. Some users will assume those rating stars are describing the products, and some will think it is about the sellers who sell them. There was this famous case reported by Nina Kollars at Defcon (watch it here; you won’t regret it) where people thought those ranking stars were seller ratings.
Your competition might spam you with fake reviews to kill your CRs right before the Black Friday event. Your content managers will miss their targets, and some will ‘graduate’ to find another job. We have seen it happen before.
Suppose you operate globally, and some of your products are successful among the Turkish minority in Germany. In that case, some German speakers might be put off by a list of reviews written in a foreign language.
It is a simple and somewhat dull example. Contingency plans are in order, scenario planning of sorts.
I have been seeing it in my practice for many years now, in many areas of product management.
We lack people with developed thinking in systems.
Those who see those counteracting/balancing loops.
Or we prefer to ignore them as this thinking is costly and, honestly, above our pay grade. Who cares that the whole thing returns through other holes in 2-3 years? The tenure doesn’t last that long, so before we know it - there is a guy who replaced that guy from before (reminds me of presidents who serve some ~ 4 years).
Battling this is notoriously hard because we have to switch off our reactive thinking (aaa, the graph goes down!) to something more advanced (if we press this red button, it goes boom here and there, and in 10 years, no one lives in that place).
Recently, I allowed myself a book-buying frenzy by acquiring books written by Senge, Sterman and a couple of others (links above). These people pushed me into unknown horizons I didn’t know existed (especially Sterman).
Godammit, we were never taught to think in wide horizons! Those who do, end up moving the needle the way they want it, succeeding in fulfilling their strategies.
I can’t say I am good at it, far from it. But I am learning.
You know, Peter Senge (one of the authors) mentioned our language as the primary inhibitor of this type of thinking. The majority of our languages are exceptionally linear. That is, we go from A to B in everything we think, say or write.
Reactiveness in its purest form.
There is no web of loops between the sentences, like sentence “15” 2 chapters before could update our chapter 4 by adding a different context. Like those “choose your own adventure” books some of us haven’t even seen. Non-linear that changes something on the pages you have already read.
The linearity is deeply ingrained in us, and I am not sure it is reversible even in the slightest.
Remember how we learned to walk? One step, gravity puts us down, so we change the balance and take another one.
Action > reaction > action. Reactive.
That’s how humans live.
Wittgenstein underlined that our language is an operating system that dictates our will and the scenarios we can imagine (and live).
I deviated a lot from the story to pour some light onto the problem.
As a Community, we have to promote system thinking in our produce. Bouncing back issues isn’t bringing us far, and those of us who managed to reach the highpoint (CPOs) have to learn the skill by running experiments, modelling those sophisticated interactions between the touchpoints, and running scenario analysis every time they do their planning.
At times, the simplest leveller shatters everything or fixes the leaky ship we are in.
In this case, we face a specific issue of coordination costs. We need it handled well.
From my experience, there are two main setups of the product function in the company:
Independent and skilled PMs under some liberal control apparatus (CPO or else). In this scenario, a company hires the best PMs and tends to save a bit on the coordination leader; not much is expected, so I have seen hired ‘garrisons’ as Jocko calls them, people good for presentations and organizing meetings. They are pretty much worthless in important domains but also relatively unharmful to the PMs at work. They usually come from huge enterprises where it is easy to hide their incompetence, but it doesn't matter as the PMs are good. If something terrible happens, people leave or stop doing what they did well - the ‘garrisons’ won’t have a clue how to fix that; they have never experienced hardships with a constant stream of decisions to be made, life-changing decisions.
The opposite spectrum is a situation where you have strong coordinators fulfilling their integrative functions, enabling mediocre product folks to craft some magic. Those hired guns are expensive, but they can keep subpar product apparatus at bay. There won’t be any sparks coming, but the plant won’t stop working. Moreover, with marketing-led / sales-led organisations, PMs are downgraded to POs, busy Jira monkeys, as we call them. No quality folks are needed here; replaceable staff.
Let’s broaden our horizons, dear colleagues; given the demographics of this world, a lot of us will have to wear more than one hat for many more years than our parents.
And the systems approach needs to be taught.
p.s. for those of you willing to enter the realm of introductory system thinking, I’d recommend starting with Sterman’s expensive but irreplaceable work.
Lastly, please check out this article; it is contextually related to this post.