TJ and I on my surprise b’day party

Acquired by Amazon— what we learned at Pillpack

I never had a doubt, from that first coffee with ✌TJ Parker and Elliot Cohen at Voltage in the summer of 2013, that @Pillpack was going to be great. That’s not to say that I expected we’d build a company from zero to $1BN in 5 years, or build one that could prove to be transformative to healthcare in the coming years, but I was smitten. So smitten in fact that I invested out of two different firms, Accomplice and Accel.

When the Amazon acquisition was announced (for a reported $1BN), I was hiking off grid with my kid in New Hampshire. So instead of writing a companion piece to Ten Things I Learned from Zoopla and doing some newsjacking like my most excellent friend David Frankel, I decided to take my time and calmly reflect with the help of management about the lessons we learned.

I hope you will like and share.


  1. Pillpack always maintained maniacal focus on the end user

You may think you have focus on your consumer, but it’s not until you work with a team like Pillpack’s that you really understand what true end-user focus means. The team overinvested in UX and design from the very early days. They never let a single piece of content or front-end go out that wasn’t pixel perfect. And it wasn’t just design. The whole flow, from first touch to CS, was pored over endlessly. Pillpack focused hard on the end-user and only the end-user and refused any distractions until it got the core customer value proposition absolutely right.

We chose a customer that had a deep and emotional need and we stayed focused on that from start to finish. I really believe it is the only key decision we made that was both necessary and sufficient for our success — Elliot

2. We decided to build full stack — PharmacyOS

The back-end of a large pharmacy is horrendous. There are countless dependencies to systems that no engineer in their right mind wants to integrate with. It’s generally a giant, antiquated mess. With lots of faxes.

Early on, Pillpack took a seemingly risky decision: the team would build a full-stack pharmacy that would free the company from the shackles of legacy and allow it to control the customer journey end-to-end. It wasn’t an easy decision to make as that meant doubling the engineering team overnight whilst revenue was still early.

Be willing to devote tons of resources and energy to the hard work of building systems, back-end-tooling, etc. rather than just building things that are “cool” — TJ

Elliot and I coined the term “PharmacyOS” during a long-walk along the Somerville community path. We asked ourselves a single question: what would it take to fully control on our own destiny? The answer: c. $10M and 25 engineers. This investment proved to be decisive.

Pharmacy / adherence isn’t a single threaded problem. You have to redesign / reinvent the whole end-to-end experience. You cannot solve it with a single piece of technology. But technology for technology sake isn’t interesting — building great experiences for customers is valuable and you can actually build a great experience in healthcare. Better experience leads to better adherence leads to better outcomes. Remove the friction! — TJ

I strongly believe this lesson is generally applicable. At Zoopla, the team did exactly the same, building the full stack down to our own billing system. At Deliveroo, ditto.

Own your stack.

3. Pillpack stood its ground against the industry giants

Healthcare is not a game for the faint of heart. When your competition is massive vertically integrated players on whom you depend for a core part of your operations and with whom you compete directly on the customer side, things can get uncomfortable, culminating in this case in a public clash with ESI.

Management never blinked and the company lived to fight another day.

If you build the right thing for customers you will ultimately prevail — TJ

Pillpack didn’t fit neatly into an established box, as we mixed components of retail, mail-order and long-term care. The team constantly had to go back and explain the need for innovation to solve the healthcare issue and how best to bring it to market. I learned about the need and value of lobbying, and TJ was extremely forward-looking in bringing some discrete but effective heavy hitters early.


  1. You can’t really grow and optimize a business at the same time.

There’s this notion in startup land that you can somehow tweak your way to success at the same time as you supercharge growth. I’ve lived through enough of these now to know this is not true. No matter how carefully you prepare for scale, every large business I have been involved with goes through a period of re-tooling and re-engineering as it absorbs the lessons learned along the way. I lived through this at PillPack, Deliveroo or Integral Ad Science. It’s a frustrating and anxious time when your burn keeps going up and you’re not seeing the progress you’d like to see on the topline. It feels like you’re taking two steps back and moving sideways right as your capital needs increase.

We had to grow, then optimize, then grow.
You have to design & do things that will scale. There were things that we were too precious about in the beginning: certain customer care interactions, the way did onboarding. We head to redesign those processes to 10x, 100x the output — Colin Raney, CMO

The thing is — it’s hard to circumvent this issue. You shouldn’t commit to code processes you don’t fully understand, and you shouldn’t industrialise before you’ve really understood what you’re solving for.

I consider myself data driven and analytical… but, metrics matter even more than you think. People think that its critical that the measurement is automated and the metric is truly “right”. However, this usually gets in the way of creating good wholistic and simple metrics. My advice is to start from the outside and work from abstract metrics to more detailed ones. Geoff [Swindle, CRO] did this exceptionally well in Salt Lake City. They started with a ton of manually generated metrics that were probably wrong. However, they got the team accustomed to metrics from day 1. They also enabled his team to prototype a ton of different metrics. As they solidified on the right ones then they started to automate them and make the measurement more accurate — Elliot.

2. Get your core team right.

Sustainable success starts at the top. You can never spend enough time choosing the right co-founders or hiring your key execs. You know this to be true, so do it right.

Culture, culture, culture. Focus on winning. Have the curiosity to learn. Maintain a founders mentality well beyond the founding team — Geoff Swindle, CRO.

I think the board at Pillpack made a real difference. David Frankel and I felt like we were part of the organism called Pillpack from day one. When we welcomed another “conviction investor” to the board, namely George Zachary at CRV, the dynamics did not change.

Having a great bod matters a ton — great board members point out issues & save time. Bad board members drain energy & waste time — TJ

But then, of course he would say that ;-)

3. Culture will see you through. Invest in it.

There’s a lot that’s been written about the value of having an explicit culture, and it’s easy to dismiss this as fluff, which frankly too many VC’s do. But when Peter Drucker wrote that “culture eats strategy for breakfast”, he was 100% right.

It’s just very hard to build a great team. The job never ends. It should be 90% of your focus IMO to recruit that team and then make sure they work well together. The other 10% should be setting direction for that team. If you have the wrong team it’s flipped. You spend 90% setting direction which only leaves you 10% to find a better team and get them to work together. It’s a vicious cycle that you have to break out of — Elliot

Culture is not just an employee handbook. It’s something you nurture and maintain. When you’re thirty people, it’s easy. When you’re three hundred and growing fast, not so much. It’s the job of management to constantly reweave the tapestry of values that makes the company unique as it grows. It’s not easy, but it’s what got us to where we are.


  1. Some businesses cannot be built without a human touch.

In the back of our mind we (the investors) had this idea that Pillpack could build a fully e-com / tech-enabled pharmacy with no human touch. Wrong.

In techland we have a tendency to think there is an answer in code to every problem in the world but in the case of Pillpack, which requires trust, it was clear we should have started with a people first approach from day one.

On a related note, I think we collectively suffered from a slight delusion that the consumer would immediately see the value of our approach. Turns out, educating consumers about a new pharmacy product is not so easy. PillPack was a new concept to market that made logical sense but developing tools to educate consumers quickly was challenging.

2. Operations are HARD.

From the outside you’d be forgiven to think Pillpack is a gimmick involving packaged meds in bags and a pretty IDEO-designed container.

Behind the scenes it’s a beast. Pillpack have pick-and-pack robots, high-speed photography for quality control, logistics, systems to transfer prescriptions from multiple pharmacies, chase refills from a variety of doctors for each patient, integrations with the payers, demand forecasting systems for meds etc etc etc. The list goes on.

I knew from the start it was going to be difficult to nail all this, but I don’t think any of us quite understood how much complex flow optimisation the company ended up having to do. This was compounded by a dual HQ set-up (Boston and Salt Lake City) and distributed pharmacy infrastructure. You really need to apply systems thinking.

The key lesson here was that Pillpack didn’t hire specialist management early enough.

Operations is super hard. Hire strong leaders here early — TJ
I think what’s particularly tough is that you are trying to find excellent people for positions you know almost nothing about.

As usual with the type of issues, flexibility and speed of execution/iteration were key to success.

We just weren’t that dogmatic about what the “product” was. We therefore didn’t have a ton of strongly held beliefs or insights that could turn out wrong. Instead we got to market as rapidly as we could and started getting feedback which guided everything else.
My best advice is the same as with most things. Have a bias for action — Elliot

3. We didn’t hire a strong CFO / COO

Ah, Yvonne Hao. Thank God for Yvonne. We all wish she’d joined us two year earlier. Finance chops, operational chops, gold dust.

Nuff said.

Thank you Jacquie, Emily, Chris, TJ, Elliot, Colin, Brian, Geoff, Yvonne, Shawn, Lenny, Siena, John 1, John2, Jordan, Gavin, Matt 1, Matt 2, Teague, Tushar, Alan and the many others I don’t know or am forgetting who helped make this a success. So proud, super happy and a little bit sad :-)

Liked it? Share + follow me on twitter @fdestin

Team Pillpack : Geoff, TJ, Yvonne, Colin and Elliot