Chef’s table

I found Netflix Chef’s Table series really inspiring. I watched all 6 episode in like 3 days. Here’s the trailer:

I loved how each chef was constantly pushing their limits, constantly learning, constantly creating. There’s no room for stagnation at the top. Not as a chef, not as an artists, not as a designer.


Slack’s Design

Loved this article about the design of Slack by Andrew Wilkinson of MetaLab:

Some highlights:

We did the logo, the marketing site, and the web and mobile apps, all in just six weeks from start to finish.

I felt the problem had already been solved. It was a crowded market and knew it would be difficult to make his product stand out from the crowd.

Most enterprise software looks like a cheap 70’s prom suit — muted blues and greys everywhere

It Feels Different

Throughout the entire product, everything seems to playfully jump around and pop off the screen.

Like a well-built home, great software focuses on giving its users hundreds of small, satisfying interactions.

In Slack, every piece of copy is seen as an opportunity to be playful.

With Slack, a bubbly, bright UI, delightful interactions, and hilarious copywriting come together to create a personality. A personality which has triggered something powerful in its users: they care about it. They want to share it with others. It feels like a favorite co-worker, not a tool or utility.

Inside Automattic’s remote hiring process

How can Automattic consistently hire the best people without ever having a single voice conversation?

Let’s face it, hiring is tough. When you ask startup founders what their biggest challenge is, hiring is one of the most common answers you’ll hear.

It’s hard enough when you have an office and can interview an applicant in-person. Automattic is 100% distributed. We hire people from all around the world. We can’t meet people in-person to interview them.

You may be wondering… how do you consistently hire great people at a distributed company?

  • Do you fly people out to your location for interviews?
  • Do you schedule phone calls at 4:00am to make it convenient for people applying from Western Australia?
  • How do you know if people will be a good culture fit?

Here’s the inside scoop…

Our hiring process is a bit different than most companies

We don’t schedule chats, we don’t fly people out, and we rarely even have a single voice call before people are hired.

You might think, “That’s crazy, there’s no way that can be effective”.

Well, here are some numbers:

I handle all design and growth hiring at Automattic. I began hiring about a year and a half ago. In that time I’ve reviewed a total of 251 resumes.

63 of those have gone on to an interview (25% of applications received)

41 have been given a trial (65% of those who I interview)

15 have gone on to a final interview (37% of those who do a trial)

14 of those have been hired (93% of those who receive a final interview)

None of those whom I’ve hired have been fired and zero have left the company voluntarily. (100% retention so far)

How do we do it?

I’ve laid out our entire process below in five easy steps:

Step 1) The pre-screen

Ultimately, much of your success with hiring will come down to the degree to which your CEO see’s it as a priority (hint: hiring should be a top priority for every CEO). Matt Mullenweg, Automattic’s founder and CEO consistently spends about 20-30% of his time on hiring. To this day, Matt pre-screens every single resume that comes into Automattic. This takes a significant portion of his time (we get lots of resumes).

“That’s insane”, you might say, “Why invest so much of the CEO’s time just pre-screening resumes? Shouldn’t this be delegated?”

No! Here’s why:

Quality control: By reviewing all applications himself upfront, Matt can ensure from the very beginning that each applicant will be a good fit for the company.
Position throttling: At any point in time, Matt knows better than anyone else which roles in the company have the greatest need for additional resources. By pre-screening all applications he can easily throttle how many applications get forwarded on to each hiring lead, thus effectively increasing or decreasing volume for every role in real time.
Pre-approval: As a hiring lead, knowing that Matt has already pre-approved all resumes that come my way helps me move forward confidently with each applicant. I never have to second-guess as to what Matt might think about a particular applicant, because he’s essentially already pre-approved every application that I see.

Step 2) Second tier resume review

We now have 315 people at Automattic. As such, we have multiple hiring leads for various roles. Every week Matt sends each hiring lead a batch of applications. Mine will consist of design and growth applications exclusively. I’ll then review, and respond to each applicant.

Note: Replying to applicants takes top priority over every other responsibility I have.

If someone doesn’t look like they’d be a good fit, I’ll typically send them the following:

Hi John,

Thanks for your application to Automattic. We don’t think there’s a great fit at this very moment, but I encourage you to keep an eye on the “Work With Us” page on, and also keep us updated and consider re-applying as your skills and contributions to open source projects grow and expand.


Dave Martin

We try to reply to everyone, whether they look like a good fit yet or not. There are good reasons for this. Not only is it the right thing to do, we’ve actually had quite a few Automatticians get hired after reapplying a second or third time.

When someone does look promising, I’ll send the following:

Hi John,

Please add me on Skype (redacted is my username). I’d love to chat with you about the position.



Typically the applicant will add me on Skype within the next 24hrs.

Step 3) Interview

Immediately after I accept someones Skype invite I’ll kick things off by saying something to the effect of:

Hi John!

Rather than set up a time to chat, let’s just keep a conversation going in Skype. I’ll ask questions, and you can answer as you have time. Could be today, tomorrow, whenever you’re available. Sound good?

Then I just wait for them to respond. By keeping things 100% text, and 100% asynchronous, I don’t have to worry about trying to schedule a day/time to chat (which can definitely be a pain across timezones).

Not only that, the concept of answering questions asynchronously is a relief for most people. Most everyone who applies is currently working elsewhere. Allowing them to answer questions whenever they’re available gives them the freedom to answer before or after work, or even on the weekend.

It also frees me up to work on other things in the meantime. I can have up to five interviews going at any one time without any difficulty, while also working on other stuff in between. It ends up working out really, really well.

After they reply to that first question I’ll let them know that we’ll just keep the chat text-only, and I’ll typically ask an open ended question. Something like:

I’d love to learn more about you. Let’s just keep it as Skype text for now (most of our communication at Automattic is done via text). Tell me a bit more about yourself. What are you passionate about? What do you enjoy doing?

I’m able to glean quite a lot out of this first question. Their answer here shows me how well they communicate. It gives me a glimpse into their interests, which helps determine whether they’ll be a good culture fit at Automattic. It also serves as a bit of an ice-breaker before we dive into questions related to the role.

After we’ve chatted about interests, I’ll dive right into questions related to the job. My goal is to keep interviews as short as possible, while at the same time extracting as much insight as possible. I’m constantly fine tuning the questions that I ask. Here are a couple of examples:

– Can you break down how comfortable are you with the following, on a scale from 1-10 (10 being expert): PHP, JS, react, jQuery, HTML, CSS?
– What one single design that you can share a link or screenshot to are you currently most proud of creating?
– When it comes to growth, would you say you skew more towards A) Doing a little bit of research, and then just starting to test stuff, or B) Devoting more time doing research up front before you narrow in on which tests to run?
– Take a product like Let’s say you were tasked with increasing engagement (which we define as the number of people who log back in each month). What are some specific ways that you might go about trying to do that?

Every question I ask is strategic. There’s no sense in wasting your time, or theirs asking needless questions. Every question you ask should deliver some level of unique insight. If it doesn’t, you should drop it, or tweak it.

On that note, if half way through the interview you’ve already made up your mind about an applicant, you might as well just end the interview there. When an interview doesn’t work out I’ll typically say something like:

Thanks so much for the interview. I appreciate your taking the time to chat. There were a number of things that I liked about your application, unfortunately I don’t think there’s a great fit at this very moment. I Would encourage you to keep an eye on the “Work With Us” page on, and also keep us updated and consider re-applying as your skills and contributions to open source projects grow and expand. All the best!

If the interview goes well, I’ll ask if they have any questions. Some applicants have lots of questions, some have just a few. The question I get asked most frequently is about next steps, to which I’ll reply:

Everyone that gets hired at Automattic goes through a paid trial process. You are given a project, which you can work on as you have time (could be an hour or two a night, could be on the weekend, whatever works for you). You’ll keep track of the hours you work, and invoice us at the end of your trial. Most people do this while still employed. You can expect the trial to last about a month (but it really depends on how much time you can put in). There is no deadline, it’s done when it’s done. How does that sound?

Some applicants will want to start right away. Some may want to clear up a block of time, and may wish to hold off for a week or two. We’ll happily accommodate whatever works best for them.

Step 4) Trials

Trial projects are essential to effective remote hiring. I can’t imagine hiring people without them.

Every single person who applies to Automattic (no matter the role) does a trial. We pay all trials the same amount ($25/hr).

Trials are a win/win solution for everyone involved.

For applicants:
– A trial gives them a chance to test out whether they’ll enjoy working remotely.
– It gives them an opportunity to take Automattic for a spin, with no strings attached.
– They can stop their trial at any time if they choose.
– Whether they end up getting hired or not, they still get paid for their time.

Trial projects are great for us for a number of reasons:

– They give us insight into how a person actually works.
– We get to see how well a person communicates over the course of a project, and how well they respond to feedback.
– Our trial projects are real projects. They consist of actual work that we need done. In that sense, we’re also getting additional stuff done in the process.

To transition from the interview into the trial itself I’ll usually tell the person:

I’ll be in touch shortly with 2 things:

1) A trial contract (just sign it and email it back)
2) Access to a private blog where we’ll communicate throughout the rest of your trial.

We’ve got an internal tool that I can enter an applicants name, email and position into, which will then send out a trial contract to the applicant.

Once I fire off a contract, I’ll post a comment on our internal hiring P2 to give HR a heads up that a trial contract will be incoming shortly.

NOTE: P2 is a WordPress theme that makes threaded discussions incredibly simple. We use P2 for a large portion of our internal communication.

I’ll then create a new private blog and invite the applicant to be an editor. This is where I’ll post the project brief. This is also where I’ll move the conversation for the remainder of the trial.

“Project brief” you say? “What does that look like?”

Have a look, here’s an example:

Project Brief

Improve the homepage design for


We’ve been meaning to give a refresh for some time now. The current design has been there for a number of years. I’d love for you to spend some time coming up with a fresh new look. Just focus on the homepage for this trial project (no need to spend time redesigning any of the other pages).


  • Automattic logo can be found here.
  • I’m happy to hop on a call and go over any of this if you’d like.
  • You can always ping me with questions here whenever you need help, or when you are ready for feedback.
  • Additional internal resources are given here

Let me know if you need access to any additional resources.


Along the way, please post all of your deliverables to this P2.

Round 1: Summary of your thoughts

I’d like for you to first summarize (just a simple bullet list is fine) the problems and opportunities you see with regard to the current website. I’ll review this to make sure we’re on the same page, and then you’ll move on to round 2.

Round 2: Rough concepts

Use whatever method you prefer to convey quick concepts (sketches, wireframes, balsamiq, etc…). Please don’t spend any time on polish at this point. I’m just looking for rough concepts. You’ll then iterate quickly based on feedback until we’re both happy with the direction.

Note: I’m not looking for any sort of re-branding. Feel free to move the logo around, and improve the layout, but no need to redesign the logo for this specific project.

Round 3: Add some polish

Get the design as close to pixel perfect as you can (or optionally hop straight to code, if you’re more comfortable doing that).

Round 4: Code it up

Last, I’ll ask that you code it up. Simple HTML and CSS is fine. Feel free to mix some JS in where appropriate if you feel comfortable doing so (but it’s in no way required).

Your final deliverable should be a zip file with all of your source files, as well as your HTML homepage.

Good luck!

There is no deadline for this project, but I encourage you to communicate early and often (not just when you’ve got something that’s pixel perfect).

I’ll touch base with the person one more time to make sure that they got the contract and access to the private blog, then I’ll just leave it in their hands.

I’ll then work with the person throughout each round of the trial project. We’ll communicate back and forth about the project via the private blog.

If things don’t end up working out, I’ll do my best to highlight why. At this point the applicant has invested quite a bit of time. I try and be as specific as possible as to why they are not going to proceed to a final interview.

If the trial goes well, I’ll chat with them about the final interview with Matt.

Step 5) Final interview

In the final interview Matt will chat with each applicant. This serves as a good opportunity to vet each applicant one last time. This is also where Matt will chat with the person about compensation, and answer any additional questions they may have.

If the final interview goes well, they are sent an offer letter.

Recap & final thoughts

  • As the CEO, you need to take hiring seriously. It should be one of your top priorities. Bookending the hiring process as the first & last touch point is an excellent way to ensure that only the best people get hired.
  • I’d love to be able to give everyone the opportunity to work at Automattic. Telling someone that they aren’t a good fit is hard, and unfortunately it doesn’t seem to get easier with time. The best thing you can do is to try and be helpful and kind to everyone, and to try to get back to people ASAP.
  • Telling people no is hard, but mistakenly bringing on the wrong people can be much worse. While you want to always be kind, and helpful to all applicants, your primary responsibility when hiring is to ensure that only the best people get hired. That is priority number one.
  • Trust your gut. You may often be tempted to “give this person a try” even though deep down you don’t feel right about it. In my experience, this rarely works out, and just ends up wasting a lot of peoples time. Two rules that I’ve learned to live by: A) never interview anyone that I’m not confident will make it to a trial, and B) never offer a trial to anyone who I’m not fairly confident will make it to a final interview.

Interested in learning more about our process? You can read some of Matt’s thoughts in this HBR article.

We’re hiring

We currently have 13 open positions. We’d love to hear from you!

MicroConf 2015

Earlier this week I attended MicroConf for the first time. What a great conference. I admittedly felt like a bit of an imposter (since I’m not actively trying to start my own company), but I still picked up a number of great insights, and met a number of interesting people.

One thing that became quickly obvious is that no one experiments with growth at the speed and efficiency that bootstrap founders experiment with growth. :-) This is of course out of necessity.

A couple of additional thoughts that resonated with me:


  • Duolingo has really stellar onboarding, as does Slack.
  •, Drip, and Wistia all offer good examples for how to run successful drip email campaigns.


Loved these insights from @Copyhackers talk:

  • There are very little actual best practices when it comes to copy. DO NOT try and copy and paste from other peoples polished successes – it never works.
  • Headlines and buttons work well when optimized together.
  • Repeat your headline copy in your button.
  • Don’t imply work is involved in button copy.
  • Every element on your page has one job. A headline is meant to keep you on the page. A button is meant to be clicked. Look at each element to see whether they are doing their one job. If not, focus your attention there.
  • In the end, no one has a clue what you should be saying to your customers (nope, not even you!). Just try a bunch of stuff and use whatever works best.
  • While you can’t copy and paste successful copy, you can mine ideas for great copy from testimonials, support emails, reviews, message boards, book reviews on Amazon. Anywhere regular people gather to talk in an un-polished way about your industry.


Some great insights from @steli:

  • Sales is about you being the most decisive person in the room.
  • Start by qualifying people – what are their needs? People just want to know what’s in it for them.
  • Follow up until you get a yes, or a no. Silence does not equal rejection. Just keep following up.
  • The magic is in the follow up. After 2-3 follow ups most people quit.
  • Keep demos short – 10 min max – focus specifically on how your product will solve their pain, then stop talking.


You always hear that you should price your products based on value. When you first launch something, the actual value your product delivers can be quite low. @robwalling talked about “aspirational pricing” where you set a price (that’s potentially too high to start off with), and you keep adding value until the value in your product eventually matches your price.


What a great community. I loved the sense of inclusion by everyone. I loved that everyone was both a teacher and a student. I loved hearing all of the inside tales about what people are working on, what’s working for them, and what isn’t. In the end, I’d highly recommend this conference.

Habit Summit Notes

Had the opportunity to attend Habit Summit this week. Here are a few notes:

Roger Dooley ( – Habit, persuasion, and neuromarketing
– Predicts use of wearables to be used to gauge biometrics in user testing studies (interesting thought)

Kintan Brahmbhatt ( – Amazon Music (reducing friction to increase engagement)
– Lamborghini story. Got remote control car as kid – no batteries. Friction killed the experience.
– Friction stops users from completing their task.
– It’s up to us as product designers to identify that friction, and reduce or eliminate it.
– Dining experience – great food, fantastic food – but check came 20min late. Just remembered that the check was late.
– Moments of disappointment have more weight on recall
– Reducing friction helps make a product more usable
– What causes friction:

1) avoidable effort
– FireTV example – voice search for TV & movie programming vs. typing in on awkward alphabetical (vs. QWERTY) on screen keyboard.

2) anticipate and avoid unnatural context switching
– Example where while watching a movie or TV show, you can pause the show, and see which actors are on the screen.
– Example where you’re watching a TV episode, you can pause the screen, tap on a character, and read the back story of that specific user up to that point in the series.
– Example, dictionary definitions built into Kindle, but also have a setting called “word-wise” where explanations for complex terms will be automatically be shown (for beginning readers).

3) Remove need to make choices
– Example with Amazon is 1-click
– Example – Amazon Fresh – makes it easy to select from past purchases

Bo Ren (Instagram) – Designing products for behavioral change
– How people see themselves – comparing to others can be extremely motivational.
– Example of door hanger experiment where they went around to neighborhoods and hung 4 diff door flyers. 1 based on saving money, one on environment, one on civic duty, one on comparing you to neighbors. The one comparing to neighbors was the only one that caused a decrease in energy usage (by 6%).

Ximena Vengoechea (Twitter) – designing people first
– Getting to know your users
– Please test assumptions before you build, let alone ship a product
– Switch from product first, to people first design
– Strategies to think like regular people, not product people:

1) Identify feelings
– What emotions are you designing for?
– How do they feel when they open my app?
– How do they feel when they close it?
– How do they feel about your competitors
– Way to test – Home screen test (what emotion do you feel)
– if you can identify the feeling, you can identify the relief

2) Consider context
– What’s my users current context?
– You have to consider urgency, state of mind
– Use whatever data you can to help inform context – get creative

3) Understand motivations
– What is my users happiness dependent upon?
– Then design a path to that.
– introvert vs extrovert

4) Get to know their values
– What is my users world view?

Ryan Hoover (Product Hunt) – The building blocks of community
– lot’s of good ideas in his slides
– 3 stages of community:

1) Friends, and friends of friends
– Good place to start a community
– Start with the people you’re most connected to

2) downhill effect – collection of people around similar interest
– Product hunt is that
– Hacker News
– Dribbble

3) platform
– Reddit

Nathalie Nahai (author) – 5 tricks to Persuasive Product Design
1) Endowed progress
– Start funnel progress off on 1, not 0
– Example carwash – Card passed out with 10 circles, and 2 stamped vs 8 circles no stamps – One with 2 stamps had much higher usage rate.
– Increases conversions
– Decreases completion time
– Losses are felt greater than gains (no matter how small the loss)

2) Sunk costs fallacy
– Once we put effort in, we hate giving up on that effort. We feel compelled to continue

3) Appointment dynamic
– They’ve spent time/invested (sunk cost), now schedule a time for them to come back
– Have to know what motivates your users
– Give them relevant initial progress
– use to create a path for them to follow

4) Opportunity cost
– Three main costs
A) Attention – acceptance, belonging, social validation (why Facebook and twitter are so successful)
B) Time – alone, bored/killing time, when waiting
C) Money – fun pain – make it hard, make the time it takes to complete great – then make it easy for them to pay to skip all of that. Adding a intermediate currency, cloaking the perceived cost. Allow people to earn currency. Offer bulk discounts

John Egan (Pinterest) – Measuring for engagement
– R&R team (retention & resurrection)

1) Activation
– Communicating the value of your product is key
– Retained MAU’s
– 30 day window – too long – 1d7’s percentage of users that come back at least once in the first week, high correlation with retained MAU’s and long-term retention
– cohort heat maps – shows how new user retention curve is changing over time

2) Engagement
– Finding your key engagement metric:
a) What action does a user need to take to get value form the product?
b) What is the minimum frequency that someone needs to use our product?
– Classify users into states
a) Core, casual, marginal
b) New, dormant, resurrected

3) Retention & Resurrection
– Each email must raise engagement, or get’s dropped
– look at negative signals – app deletion, spam reports, unsubscribes

Tristan Harris (How to measure success on time spent well)
– After a better web
– One with better motives – not just about money
– Organic type initiative, but for the web

Jake Knapp & Daniel Burka (Google Ventures) – 5 day idea sprint
– Story about Nintendo – Rented for 4 days from friend, and decided to was too addicting
– Startups today – MVP, launch, get feedback, iterate/pivot
– Anther approach – 5 day design sprint – focus is on learning
– Lot’s of details here:
– Super entertaining presentation

Remembering to separate proposed solutions from insights & observations

A couple weeks ago I listed out a bunch of activities designers can use to reduce time to clarity.

Upon completing one of these activities I’ll generally report back to my team to recap everything that I’ve observed. This can be a really valuable step in the design process, for a couple of reasons:

A) By itemizing everything, I’m able to better mentally process what I’ve observed.
B) Other people on my team may end up seeing patterns that I didn’t see.
C) Everyone on my team will have the chance to learn (which is really the main goal of these activities)
D) By reporting back, I can help everyone on my team better understand the thought process behind my designs.

A quick word of caution

One pitfall I commonly find myself falling into is my tendency to mix “proposed solutions” in with my recap of insights & observations.

It’s easy to do.

But doing so can easily derail everything.

Why is that?

By mixing solutions in with observations, I’m essentially mixing biased ideas in with relatively unbiased insights. This just increases the chance for:

– disagreements (the focus of these recaps should be on the observations – on what I learned – ideally this should lead to very few disagreements)
– “rabbit hole” discussions that steal the focus of the thread
– my unbiased insights being ignored, or even dismissed completely

One easy solution

Why not just separate the two out completely?

1) Post my insights and observations, being careful not to mix in any proposed solutions.
2) Then as a threaded reply to that main post, go ahead and list out any solutions that I came up with.

By compartmentalizing the two, I can help keep the primary focus of the discussion on the insights & observations. Teammates are then free to pick and choose which discussions they want to participate in, be that:

– insights and observations
– proposed solutions
– or both

I’d love to hear your thoughts.