Why We Killed A $150,000 Feature

I can’t tell if he wants to laugh, cry or punch me. His emotions right now are about the same as mine were 5 hours ago. I’ve been up for 39 hours.

My head is throbbing.

I’m unsure if it’s from the stress of this huge decision I’ve been considering through the early hours of the morning, or the 4 cans of Walmart’s own brand energy drink.

I’m shaken from my microsleep by his reaction:

“Honestly, I can’t disagree with anything you’ve just said.”

He releases a sigh that carries with it the colossus weight of the codebase I’ve just unburdened him of. I sigh, sharing his relief, and slump down into my chair.

Reminding me why I love having this guy as my equal partner, he’s immediately into action mode:

“OK, it will take me a few days to rip it out. I need to make some changes to the API, database, figure out the UI changes… But yeah, let’s delete it!”

Reading the data from every angle, reviewing every single customer conversation over the last 9 months, and spending days obsessively Googling over the competitive landscape… nothing prepares you for the decision to delete a huge feature. A decision better shared with a cofounder.

Especially a feature that’s cost you $150,000 of investor’s cash.

Here’s the journey that led us up to making this decision, the lessons I learned about being a product manager, and how it’s totally changed the way I think about building products in early stage startups:

Building Great Products Is Really, Really Hard

It’s a simple disclaimer, and one I find myself saying to other SaaS founders a lot. In my opinion, no rules, laws or frameworks can guarantee that you’ll create an amazing product that will hit $25k MRR after 6 months, $1m ARR after 18 months…. You can’t even have a guarantee you’ll hit $1k MRR!

I have read the Lean Startup. 8 Times. I’ve given workshops on the Lean Startup principles. I’ve read hundreds of posts from talented product managers.

And yet in the real world, outside of the perfect lab environment that these fool-proof frameworks are described, the overwhelming number of variables at play mean that building a new product is messy, sticky and a bit unpredictable.

What Exactly Did We Build That Was So Bad?

I’m humble enough to say Trakio’s product vision in 2013 was vague and naive. We were going to help SaaS companies with their customer lifecycle marketing (I’m a big fan of Pirate Metrics) in some kind of way, because we had decided to build this cool customer analytics database, and it seemed like a good thing to do with the tech.

As part of our vision to solve “Every and any problem a subscription company might have…” we thought that selling Trakio would be easier if we added a marketing automation and mini CRM feature onto our basic analytics and metrics features.

After all, a few of our early customers had said they thought it was a great idea. It would save them money on having to buy a separate emailing tool!

This means our tool would now be better value – because customers would get all of these extra features included – so we’ll sell more subscriptions! More features means a better product!

Of course, I now know this is bullshit.

We made 3 big mistakes here.

Mistake #1 – More Features Makes A Product A Lot Harder To Sell

Dropbox – we store your files.

Box – we store your files, and have collaboration, and have these cool business controls, and special enterprise compliance options, and these services to help get setup, and…

Box has had a long and painful struggle, and is valued at $1.7Bn. Dropbox has had life pretty easy, and is valued at $10Bn.

Simple is easy to sell. And I learned this the hard way on sales call after sales call. (And just as painfully, watching the conversion rates tank in our analytics).

When we’d finished playing with it, our new “we do a bit of everything” landing page dropped to <1% conversion on signup. Previously, we’d seen as high as 7%.

Our 1 liner pitch at meetups became messy. Investors stopped reaching out to us cold (a strange, but in my opinion, satisfyingly correlated metric of how obvious your marketing messaging is)

We started to go weeks without a single new paying user, and I started to seriously lose faith in pitching the product to anyone.

Mistake #2 – More Features Makes It A Lot Harder To Build Everything To A Great Standard

I think in todays startup ecosystem, it’s too easy to forget that writing amazing bug free code, and making sure it has an amazing design and UI, is really difficult.

We’ve become obsessed with “MVP”. This idea of minimal. Ship a feature in 4 hours. Next card. Apply Twitter bootstrap UI. Next card.

We create a load of half-assed features. And these half-assed features build up.

I’m sure you justify that you’ll “Double down on the ones that get traction!”. But that never happens. Because your engineers and designers are too busy being pushed into shipping the next new feature.

If you double your features, your product quality will probably halve.

Many founders who I’ve said this to have disagreed with me.

  • “Double the team size”
  • “Hire more senior coders”
  • “Just hire an office of developers in India!”

These founders are usually ones with a lot of half-assed features, and they’re probably halfway through the process of lying to themselves about that, like Matt and I were.

Mistake #3 – Your Customers Should Not Be Dictating Your Product Strategy

“Get out of the building.” 

We’re told to go out and talk to customers. Get people to pay you before you build the product.

But what happens if you get out of the building, speak to 10 Accountants at an event and convince all 10 to pay you, but after building the tool no other accountants buy it?

You probably go back to your 10 customers and ask them what else they might like. You add more features, they pay you more money. Now you have an even bigger and more complicated product.


Taking product strategy advice from your customers can be poisonous.

Listening to your customers is recommended, for sure.

But it’s your product and your vision, not theirs. Your customers should never be telling you what to build. You decide what to build, and then test it on them.

The Costs: $150,000 And 9 Months

When we ran the numbers, we estimated we had spent $150,000 of our investor cash and revenue building these extra features, and the marketing & design that went along with them.

But much more seriously, we lost around 6-9 months of time. This opportunity cost was huge to us.

The Happy Ending?

Firstly, we have amazing investors. They understand building something awesome is a journey, and the investment they made was an investment in the team, not simply purchasing code.

Good investors know that the founders’ personal development is part of the startup process, and I think this is something that our more recent investors have recognised when we tell them our story.

Secondly, deciding that we needed to regroup on the product and re-discover our original vision (and conviction for that vision!) pushed us to apply for top tier Startup Accelerators around the world!

We were flown out by YCombinator for a whirlind experience meeting Sam Altman. We were lucky enough to be a part of the Techstars NYC application process (and meet the amazing team at IA Ventures). And we were eventually seduced by B2B focussed accelerator with close ties to Stanford, Alchemist Accelerator!

Joining an accelerator allowed us to go back to 100% product focus. We ripped everything apart. We hit delete a few more times. We looked deep into the code, and deep into ourselves, and what we found was the passion to rebuild the platform into todays Trakio!

Published by

Liam Gooding

Liam is the cofounder and CEO of Trakio. Previously an engineer, he writes about growing subscription companies using data-driven techniques and inside glimpses to Trakio's own growth journey. He wrote a book, "Growth Pirate!" which discusses data-driven growth strategies for startups.

  • flackjap

    But do you really need to delete features? Can’t you just hide them / move them in something like “Other features” sub-menu?

    Also, on a side note, IMO trak.io front page could benefit from better texts … It’s like you avoided putting up more info about product features below the fold and decided to stuff information above. At least you could cut something and still make it stand out. For example “All Of Your Customer Data Connected In One Platform”, could be shortened with “Customer data. One platform.”. And instead of bunch of information that’s right below and that tries to explain the visuals, just create simple gifs showing UI motions. And if you think that it would kill the page loading speed, use something like Sublime Text got on their homepage – https://github.com/sublimehq/anim_encoder

    • Naeem Bari

      You can hide the simple stuff, but big stuff like this article talks about, that touches a lot of your code base, you remove. If it stays it clutters up clean code, makes optimization difficult, is a vector for future bugs. You also find yourself in 2 years with junior devs working around this unknown mess. Refactor continuously – we budget 10% of our dev time to refactoring.

      • flackjap

        I see. That’s what the Technology Debt as a metaphor should also cover. It’s very easy to loose agility in terms of being able to quickly maneuver the change requests created by scaling out or adapting to market fluctuations.

        Me and my team are currently at the point of stacking up features in one of our products to be. I’m really thankful for the given insights in this blog post. Gonna cross out some features asap 🙂 Cheers!

  • Joshua_ad

    Great article! Luckily I have never had to run into this kind of problem, I know as a programmer developer that this would be heart wrenching to say the least.

    On the brightside, some of that code could be used for something else or another project, so it may not totally have been a waste or be dead.

  • John_ad

    I Still Have No Idea What Your Product Does From Looking At The Home Page, And Why You Need To Capitalize OR UPPERCASE EVERY WORD.

  • Just delete it!

  • Conrad Kleinespel

    Hi Liam. Thanks for sharing this story. It’s amazing how every company that looks/is successful seems to have had failures like the one you describe here.

    It’s inspiring to see you staying strong, to see you keep going and to see you building new products, overcoming the roadblocks as they come along.

    • Thanks Conrad! We thought it was important to share this, especially “after the fact” to show other teams that cutting features, even in a radical way as this, can be a *good* thing and we’re able to show how much stronger our product & company have become because of it.

      I totally agree with you that way more companies go through this but don’t talk about it publicly. But yeah, the reality is that product dev is hard and feature creep happens to the best of us!

  • benfavre

    You have a completely different page layout for every link in the menu, each of theses page designs changes/moves the menu options.

    Several pages repeat the exact same words.

    The whole thing looks like a mess.

    I understand it’s freeing to press delete on some technical dept.

    But i’m sorry …
    There is no easy way to see what your product does.
    I think you guys should invest some time on cleaning things up instead of running around like pitching/looking for funding, …

    Not one screenshot, no docs, no anything actually.

    Im am at a loss to why this even came up in my twitter timeline. Where am I.

    • Conrad Kleinespel

      From what I remember, Trak.io didn’t use to be “Entreprise Only”. From what I remember a while back, you could register an account easily and it seemed aimed at companies of any size.

      Would be interesting to see why the marketing has changed so much. (Or maybe its just my memories that are wrong.)

      • Hi Conrad

        Correct – while we were building this new product in stealth, we did all of our customer development with a small number of larger customers.

        If you click through from the homepage and checkout the pre-launch landing page for Metrix, you can see that our pricing is *extremely* accessible, at $10 per user per month

    • This seems an unnecessarily negative comment Ben, but thank you for reading the post.

      • benfavre

        I agree my tone is totally unacceptable and uncalled for.
        My apoligies.

        I’m sure it will all fall into place. But that was an honest reaction (not sure what it says about me …)
        The whole thing of no screenshots, no easy info on what the product really does, combined with the layout issues on the different sites makes the whole thing feel off.

      • Just Visiting

        Leaving aside the tone – Ben did flag a question -it is important to make your message as easy for your reader to grasp.
        I’m reading alot recently about making web content talk about the users’ problem that a product will solve: not about what the product does.

        I’d welcome your views on whether your target buyers (which might be me) will understand this pitch:
        “Connect Each Customer Data Source In Just A Few Clicks To Create A Single Unified View Of Each Customer. Our Beautiful iPhone App Makes It Easy For Everyone In The Organisation To Access The Data They Want, Whenever They Need It”

        • Hey – that’s a really great point to raise. Experimenting with different headlines and iterating based on the results is by far the “lowest hanging fruit” to improve conversion rates.

          Needless to say – we’re testing a LOT of different headlines and also using the “Ask Framework” method to discover the best language. Bearing in mind, we have multiple buyer personas in our audience which makes things a bit tricker!

  • nadimtuhin

    In our startup we building http://getonepager.com this plugin. I planned for a ton of features while my ceo always pushed for less features. I now know how important is it not to listen to customers and focus on one single thing. Thanks a lot for this write up 🙂

  • Just sent this as a must-read to my team – thanks for sharing the hard stuff.

    • Thats great to hear Laura – if they have any follow up questions, don’t hesitate to shoot an email to hello@trak.io

  • This is a very good read!

  • A very wicked geekie

    About point #3, I agree heartfully that customers should not, and never dictate your strategy. This said, your development in that section makes me think you’ve misread the Lean Startup. Getting out of the building is not “ask around random strangers for what to put in to get a great product”. It’s not “please fill in this letter to Santa Claus” either. It’s about listening to what their lives are with the problem that you’re trying to solve for them. Don’t ask customers what they need, they’ll think hard of something to please you (you’re a nice guy, let’s give you food for thought, why, yes, a piano tie that links to your phone’s speakers would be a fantastic thing, I’d give it around to all my friends at Christmas), instead explore their problem. Only ever discuss their problem, the hacks or work around they end up doing to solve it, and what else happens when the problem arises or the hack succeeds. And customer interviews should never, ever be a selling pitch. If you’re selling, you’re not listening. And if you’re not listening, you’re not learning what your product should achieve (rather than what it should feature). If you ask customers what they would like, they’ll tell you they want a pony, without thinking of all the manure that comes with the pony because that was not a long and hard reflection. It was a spur of the moment thought that’s weightless to them because it’s not their time, their money nor their job. The “get out of the building” part of customer development, and the “talk to humans” of Lean Startup are the hardest, scariest part of being a Lean Startup entrepreneur.

  • eugenevdm

    This is so interesting. Some time has passed. What’s the outcome? Does trakio still exist?