The FIP Journal #20: the good, the bad & the ugly feedback

Marc & Patrik
8 min read

Patrik

Wise words of the week

Your most unhappy customers are your greatest source of learning. — Bill Gates

In theory there is no difference between theory and practice, while in practice there is. — Benjamin Brewster

Journal

The other day I was joking with Marc: having to actively prioritize tasks means having too few customers.

Like I said last week, user feedback is pure gold. The good, the bad, and the ugly.

The good. A bit of reinforcement and reminding us that we are on the right track is really useful. Of course, we want to solve a problem. But people being happy about our solution means we hit a sweet spot and manage to solve the problem in both a good and pleasant way. So, when users tell us “I really like the product and it’s clear you have spent a lot of time building it”, it’s worth at least a few hours of therapy! Marc actually created a Slack channel called “love-wall” to share those compliments.

The bad. There are obviously (constructive) complaints. We want to solve a problem for everyone, not just for us. And life is so beautiful and complex that we could not think of all possible corner cases. Unmarried people, self-employed workers, people working for international organizations with tax-free income, divorces, subletting rooms, or retiring abroad. You name it and the list is quite long. This is where it helps us prioritize and understand that we are all individuals with different walks of life. Sometimes we knew but thought it was way too rare to care about, and sometimes we didn’t even know about it. Like pension plans based on the final salary rather than on the accumulated capital. I thought they were a thing of the past, until a user noted that we don’t support them.

There are also the ones who use the software in a way we didn’t expect. After a while, I know what works well and what doesn’t, so I tend to avoid modeling situations that are pushing the software over the edge. We are currently getting everyone on the waitlist into the tool and we had many new users over the weekend. Many tried to use the option that does an exhaustive search of all options. It usually works, unless the situation is complex and simulating it causes a combinatorial explosion of cases to test. Like being married, with multiple pillar 3a accounts each and multiple vested benefits accounts. I already had a fail-safe by timing out simulations after 90 minutes, but this weekend we had 6 or 7 of those. So I’m currently looking into this (more in the next section).

The ugly feedback, mostly not very constructive. Like a recent thread on Reddit saying we are greedy and asking too much for something that can be coded for 20 USD with Claude. Well, I’m sorry for those who got that impression. I’m probably spending 20+ hours per week on building the tool and managing the accounting for the company, and we haven’t paid a salary yet. As for Claude (or any AI) building something similar, well, I’m still waiting for it to happen and I would be the happiest person if it was possible. AI clearly can build a tool for doing a simulation and generate a really cute user interface. But for the logic of the simulation, we are still not there. The information is public, but as of last week, Cursor was not able to write the correct logic for computing the income tax for the canton of Graubünden. It wrote something, obviously, but it did not match reality.

How did I use AI this week?

I’m still very impressed by how well Cursor (in this case using Grok) understands our software.

I asked what could be done to improve finding an optimal way to claim the pillars. Right now the software does an exhaustive search of all possible combinations of claim dates. But it often runs into a combinatorial explosion and takes too long. I asked Grok to investigate two directions.

The tactical one: making the computation cheaper. Somehow all optimizations seem to revolve around switching away from maps to slices. The map is expensive to allocate and creates more work for the garbage collector. It’s a tradeoff, so if maps are long-lived and have hundreds of items they are a good choice. But for short-lived computations (e.g. when collecting data at the end of a month) and with only a few entries, the overhead is just too big. So I have a structure which is a slice of key-value pairs, and which supports insertion and lookup. I asked the AI to use this structure to replace maps if it assesses that the performance is better.

The strategic one: does the simulation really need to go into all possible combinations of pillar 3a claim dates? If someone has a pillar 3a account, there are 6 possible claim years. With two accounts it becomes 6 x 6 = 36 different combinations. Married? Then it’s 36 x 36 = 1'296. So this quickly adds up. Here Grok suggested just computing one person at a time and then joining the results. So it becomes closer to 36 + 36 = 72 combinations, plus a bit of fiddling.

I’m quite impressed by the suggestion, given this is already one of the most complex parts of the software. But I also want to be convinced that this is the right thing to do and that we are not missing anything or making things worse.


Marc

The build

We continued our soft launch last week (you may have received an invitation!) by doubling the number of waitlist members we emailed. That, and user research and discovery to keep prioritizing as best we can.

Sometimes you (wishfully) think that it’s just two or three clicks and done, and you can get back to improving the UX and frontend of the product. But damn, go-to-market and communication are a job in themselves (surprise haha!)

Jokes aside, the thing that did hurt the most was that one of our “pillar withdrawal” optimizations is… optimizing things too much, to the point where we got server alerts each time someone tried it.

Talk about helping us (ehmm, Patrik) prioritize what to focus our time on!

We decided to actually merge the two “pillar withdrawal” optimization types in order to solve this. And as a UX lover, I think it’s perfect as “less is more” (read: more intuitive!)

Customer voice

I just subscribed to the Snapshot option this Friday and tested the Planner thoroughly. It is a really great tool which helps me to better identify where I am currently at regarding my FI journey. — Lukas

The edge

The tension this past week was one that Product Managers know well: how to prioritize between all the user requests… arghhh!

Should we add cohabitation (vs. married) first? Oh well, yeah, but wait, the optimization speed should be our first prio! What about investment strategies, can’t we make them custom? You see what I mean. As I said to Patrik, at least I’d rather be in this mode than looking for new users!

On top of that, I was telling Patrik recently that he was putting in more hours than me, and that it didn’t feel fair to me. Until last week, when I talked about my own FI on the MP blog, and that brought many new people to FI Planner as it’s the tool that helped me make my own decision. Some would call it indirect marketing, or word-of-mouth marketing. Whatever the name, I found some peace with it as indeed, all the time I put into the MP blog (and not on FI Planner directly) also contributes to its visibility, hence its success. Because ultimately we wanna help people reach their own FI. That put me at peace with how we split the work with Patrik.

Sidenote: From the start, Patrik didn’t agree with how I saw this situation.

That’s all for the self-therapy part haha!

Note to a friend

KISS. Keep It Stupid Simple. That’s what all good products I know have in common. And that’s where I want to see FI Planner going, vs. making it a Christmas tree with bells and whistles that nobody understands.

One thing that became crystal clear to me is the pillar optimization strategy: we must at least reduce it from [None / 65 / Optimize (Good enough, very fast) / Optimize (Maximum, takes longer)] to [None / 65 / Fiscally optimized]. That’s one option out of four gone. 25% simpler!

And the same goes for our list of events. As we added use cases (cool!), it became harder to read (not cool). But that’s also how you best do product development, by crossing each bridge when you come to it. And now we are there ;)

Tool of the week

Fable 5.2. My new go-to model for complex tasks: creative ones such as a product manager partner, as well as more tactical ones such as our Journal publication workflow (including copy editing, and translations into CH-DE and FR). I used Fable 5.1 beforehand, but I feel 5.2 “gets it” more often than 5.1 did. Also it’s a bit less verbose AND costs less. Where do I sign up?! 😉

View full size