A cartoon
Credit: cartoonresource / Adobe Stock

Endless refinement of your idea. Completely lost in customer insights. Drawn in data. And not much progress apart from more and more user interviews. In essence, you’re stuck. This is how analysis paralysis feels like. The problem? You don’t usually see it until it’s already too late…

After my talk at La Product Conference in Madrid, someone in the audience asked me: “How do we avoid analysis paralysis?”.

My answer covered a couple of these points below but I decided to write them down, as I believe it can be valuable for product people out there.

So, here are 7 things you can do to stop and avoid analysis paralysis.

But before we dive into it, it’s important to be aware of what tends to cause it:

  • Perfectionism — we all want that shiny product to hit the spot when released… so, “just one more round of interviews!” can sound like an easy call. It’s usually not. Fight this mindset at all costs. Progress over perfection.
  • Fear — either related to a lack of psychological safety or a high-risk initiative. Either way, fear leads to slow decision-making and feeds perfectionism. But also tends to be correlated with not knowing how to actually de-risk an idea.
  • Misunderstanding of what great discovery looks like — because we don’t know how to effectively de-risk an idea, we tend to associate more data with better insights. So, sticking with discovery indefinitely can sound appealing from a risk reduction perspective. Unfortunately, it’s not how it works.
  • Discovery fanaticism — a typical discovery fanatic sees her/his job as to “do discovery and design”, usually following a certain framework religiously. This can influence the rest of the team to keep talking with customers endlessly, without real traction.
  • Lack of leadership — when product teams are surrounded by poor product leadership, teams don’t have goals to hit, a vision to pursue, or a strategy to follow, and are given either a “big shebang” product-to-build (which can lead to endless product validation) or a ridiculously vague opportunity (which can lead to endless discovery given the lack of context and a large number of possible paths to pursue).

If any of these sound familiar, then analysis paralysis is here to get you sooner or later, my friend. But fear not. Here’s what you can do:

1. Lead the way with hypotheses

Approaching discovery in a systematic and scientific way is a key antidote for analysis paralysis.

Avoid just going to customer interviews without a specific hypothesis to test.

We do this in generative research, which is characterized by an open and exploratory approach. It’s powerful when we want to better understand our users, customers, and dive deep into their contexts. But it often lures us to keep seeking the “truth” by just observing and enquiring all these people, again and again. The hope is that we learn so much that we’re able to generate great ideas. Although this is part of the toolkit you need to use, real progress tends to happen when we actually start testing some of these ideas.

A diagram describing unknown problems and known problems

In other words, we need to avoid being in the so-called “problem discovery” forever. Maybe a couple of days, one week at maximum

The real magic happens when we take what we know, ideate, and start testing the assumptions that underpin our idea.

That will keep you on track and help you progress.

A great way to do this is to use e.g. Strategyzer’s test cards to articulate your hypothesis (you can use other formats, as long as you’re reflecting on the key points).

A test card format
Test card from Strategyzer

2. Measure evidence and confidence levels

By breaking down big questions into smaller and specific hypotheses, and then designing experiments to test them — we can generate evidence to guide our decisions.

But how do we know we’re collecting strong enough evidence? How do we ensure that we’re all on the same page? That’s where a common framework to measure evidence and confidence levels comes into play.

Confidence levels spectrum
From Strategyzer

By evaluating our evidence using a framework that works for you, we’re able to establish a common language and a shared understanding of the strength of our evidence.

This provides clarity on what evidence really means. Is it an irrefutable truth from the market (5) or just a regular claim from a few customers (1)?

Striving for this level of consensus will speed up decision-making, and is a tool in our arsenal to combat analysis paralysis.

3. Time box your experiments

Good old Parkinson’s law: “Work expands to fill the time available for its completion”.

Not having deadlines can cause people to drag things out. Long deadlines will force people to fill their time with stuff that doesn’t really matter.

In essence, time-boxing your experiments appropriately will help you discover value sooner.

I’ve observed teams that didn’t set any deadlines for themselves when it comes to discovery and before they knew it, they were paralyzed.

Set a deadline to the best of your ability. A discovery sprint. And force yourself to learn what you need in that timeframe, acknowledging that you’ll never have all the information you need…

The point with time boxing is not for you to become a time-control freak. It’s also not to reduce the quality of your experiments just because you need to hit a tight deadline. Different experiments require more time than others. The goal is to learn, sooner. Not only from your customers but also from how you work as a team and how long it takes you to run certain experiments.

4. Think ahead: sequences of experiments

You should always design further and higher fidelity experiments based on the evidence you’ve collected from the previous ones. But chances are that you’ll be able to think ahead a little bit. Not too much… Just enough.

If you can foresee the next experiment, you can begin to evaluate if there might be obstacles in the way. I’ll tell you a little story. I was once coaching a team who concluded, at some point, that a fake landing page experiment with some social media ads to generate traffic into it would be a great next experiment.

They were not done dissecting their learnings from the previous tests, but the Product Manager started evaluating what could potentially be blocking them to run such an experiment. It turned out that Marketing was a little protective of their ads and wanted to know all sorts of things, do all the copywriting, and so on. We spent quite some time aligning and clarifying our intentions. And by the time the team was ready to design that experiment, the coast was clear.

Such small momentum wins matter to keep the ball rolling.

5. Visualize your discovery and discuss your progress

The research is pretty clear here: making our work visible improves communication, fosters collaboration, makes it easier to identify bottlenecks and potential issues, and boosts motivation as it helps visualize progress. Not surprisingly, it helps you avoid analysis paralysis. But there’s a twist. If you start visualizing everything, this may lead to information overload which may affect your ability to make decisions. In other words, keep it simple.

A kanban board for discovery
Using Kanban to visualize your experiments

I like to map things out using a Kanban board. You can create policies to help you stay on track too, such as:

  • Pull: Don’t start new experiments (or prototyping ideas) unless there’s capacity, the whole product trio decided to do that together, or until we have dissected the learnings from the previous experiments;
  • Work-in-progress (WIP) limits: no more than 2 experiments at a time, as a rule of thumb. Be very careful with testing several hypotheses at the same time;
  • Definition of Done (DoD): an experiment is done when we filled out a learning card and shared the findings across the team/organization in a very digestible and simple way;
  • Blocker: a blocker is highlighted when something is preventing progress (e.g. access to users, dependencies, etc.).
A kanban board for discovery
Using Kanban to visualize your experiments

Host regular retrospectives to remind you what you want to achieve, why, when, and what you have learned so far.

6. Review your goals

Regardless of the goals you may have or in which format (e.g. OKRs), we should revisit our progress regularly. I usually recommend an OKR check-in every two weeks. This will help us avoid paralysis as we allocate time to reflect on whether we’re moving fast enough to achieve the results we held ourselves accountable for.

7. Celebrate quick learnings and failures

Every time we prove ourselves to be wrong, we celebrate. Every time we learn something new, we celebrate. Get in the habit of celebrating small learnings and moving on fast. It’s a mindset.

8. Make faster decisions based on their reversibility

If you ever feel stuck, ask yourself:

Based on what we know now, are there still any major risks of doing X?

If you keep answering yes, despite having done a bunch of experiments (a symptom of analysis paralysis), ask yourself:

How reversible is this decision? Can we go back if it turns out to be the wrong call?

Sometimes, we just need to make the shot to the best of our ability and with the information we have at that time. Reed Hofmann calls it informed intuition. It’s naïve and, I would argue, silly to make decisions with zero evidence because “there’s no time!”. There’s always time to generate some light evidence… But, sometimes, there’s no need for high-fidelity experiments — either because the risk is low or because the decision you’re about to make is totally reversible.

In other words, if you fuc* it up, you can go back and the impact is minimal. So just make the decision, progress, learn, and have fun!


Source link