If I rewind and replay this year and take the shortest path to where Wednesday is today, I will probably add a huge chunk of time to my life. So is the nature of how things work. You need to experience the wrong decisions to realize the right ones.
Towards the end of the year, I take a few days to write my thoughts on the year that has passed. It helps me reflect on the decisions taken, whether strategic, business, or otherwise. As they say, the dots only join in hindsight.
2023 started with a bang, and it will also end with one. It’s the middle, and that’s the exciting bit. It’s always the messy middle. This year was exciting in its own way. We experienced a new wave of automation (AI) taking the main stage (at least for knowledge workers), startups saw funding dry out, and everywhere I see, there is news of layoffs. It is not the most conducive environment when running a bootstrapped business providing services to digital businesses.
If you're new to Wednesday, in a nutshell, Wednesday is a Software Design and Development agency. We've perfected frameworks that create small implementation teams that deliver exceptional results. You can learn more about us here: our work, our thoughts, how we work.
Musings on the Agency Model
The agency model is rather simple. It’s using your expertise to help people get from point A to point B in the shortest and best way possible. That’s it.
In the world of AI, however, expertise is getting commoditized, and positioning an agency is going to get tougher. Until I started Wednesday, I didn't know how difficult and interesting positioning can be for an agency. Learning about what prospects think about and how they perceive a company’s service offering is fascinating.
To understand this better, we need to go back in history. Geography has had a big role to play for an agency and probably still does, but in a different way. Back in the day, local agencies ruled their territory. It’s always easier to work with a company that’s right around the corner, isn’t it? But there was another side to the coin: agencies couldn’t specialize in a specific work area because their client base was limited. How often will a set of ~20 companies have deep tech problems for you to come aboard and help with?
As a result, most agencies were generalists. They spread themselves wide and offered an array of services. Then came the internet, and this changed everything.
Agencies across territories started infringing on clients that were loyal to a few. You could now provide services digitally and win customers by reducing the barrier to entry or price. How did everyone respond?
- We have the best people
- We have the best processes
- We save you time
- We save you money (lower rates)
- We do everything under the sun, moon, and solar system
When I think about the above from a customer’s perspective, none of it is verifiable. Apart from the agency just saying it.
The problem I’ve realized is that agency principals are too focused on what competitors are doing and less focused on what customers want. There is always an agency that they want to be like, which is usually decades ahead of them. What they did to get there was during a different era. My guess is that the path to reach the success they have will most likely be different, and that’s a problem that is worth solving for every agency principal.
A new endeavor that we started this year was market research. We opened up our calendars and started talking to people who fit our ideal customer persona. Be it engineering decision-makers, entrepreneurs, product managers, etc. Heck, I even started a podcast to give center stage to entrepreneurs to share their stories, and, in turn, I learned firsthand the details on how to craft a positioning that best suits what they are looking for in a team.
The learnings have been extremely useful in crafting my thoughts around what to do in the coming years.
- Don’t stop doing what works: If you have found something people desire or want, double down. Principals have a big tendency to look over the fence. For us, this has been application development and design. Irrespective of what AI will do.
- Expand on the moat: Do things that increase your differentiator in the areas that work for you. Our customers have preferred working with us because of the combination of great engineering skills, product thinking, and the practices the teams follow. It just gives results. We will double down on refining these practices and bringing them to the forefront.
- Laser-sharp focus: Not being influenced by something is rather difficult in today's time. Agencies that have a focus are better suited to tide through.
These learnings are a result of the experiences that I’ve had so far running Wednesday. If you’re reading this and are early in your journey and don’t know what works, my suggestion would be to first find that.
The best ideas for agencies today are not around skills. That it’s a race to the bottom. However, the best ideas will probably be at the intersection of multiple work areas. Maybe art, engineering, and product or people management and sales, etc. Look for insights at the intersection of these areas. Our sweet spot is the intersection between skills, delivery activities or practices, and product thinking. Find yours.
Musings on knowledge workers
Four years ago, if you’d asked me to jump from a plane, I would have in a beat, just for the thrill of it. I had exited a venture-funded startup and was on the moon. Everything was possible. A lot, however, changes with time. Keeping me far away from that plane now is all I can think about.
I visit my alma mater every year. I love talking to young engineers, looking over the projects they have built, the research they have done, and the papers they want to publish :). Over the years, this has changed what I expect from people and helped me manage better.
We’re highly influenced by the environment around us. It shapes our thoughts, personalities, judgments, beliefs, and much more. Talking to the next generation has made me realize:
- They have too much information but very little knowledge
- They are highly influenced by social currency (likes, follows, what have you)
- Extremely opinionated without the experiences to back it up
- and a lot more…
While I realize these statements may seem polarizing,. My opinions were formed not by talking to a few people but probably by over 100 knowledge workers just this year. Both through my visits to my alma mater and the interviews I’m a part of.
There is a popular saying: “To a man with a hammer, everything is a nail.” Reality is multi-disciplinary. If you’re highly opinionated, you’ll approach every problem with a hammer. Everything is a nail for you. Wouldn’t it be better to accumulate more tools through experiencing life and make a checklist on what to use and when? Wouldn’t that be a worthy life?
I also think that the younger generation fails to understand the importance of the environment. A lot is taken for granted, and being extremely opinionated forces one to take decisions that just don’t make any sense in hindsight. A popular one that I’ve come across this year has been regarding the reasons to quit.
Note: Before you read further, I want to mention that it's okay to quit your job as long as the reasons are sound to you and not something you would regret in the future.
Working within constraints is the very nature of growth. A plant growing in isolation doesn’t create a forest. With constraints, there has to be feedback and the ability to actively listen. Not being able to listen to valid arguments will lead to a downward spiral. You’re better off in an environment where you grow consistently rather than having triggers of exuberance. It took Michael Jordan 7 seasons to win his first championship, but take him out of basketball, and no one remembers his baseball career. The environment trumps all. If you’ve found a good one, stick to it, accumulate tools through experience, and actively listen. In the long run, it pays off.
The year that was, at Wednesday
2023 was a year filled with experiments, from marketing, sales, technology, design, delivery, etc.
We love experiments here at Wednesday. But we're cognisant of our time. We know when to stop doing something; this has been one of our greatest strengths. We can say, "Hey, I tried this, and it didn't work." We don't get attached to our ideas. We're also good at doubling down on something that has worked.
One of our goals this year was to focus heavily on building systems that allow us to measure software development quality. Why, you ask? We routinely are in a situation where we have excess work. In cases like these, we like to work with external partners. I spoke to at least 20 service providers and realized everyone says the same thing. We have the best quality, best experience, best teams, etc. None of this is verifiable externally.
We have incorporated DORA metrics measurement and reporting across all Launch, Amplify, & Control projects. DORA is a set of measurements that teams can use to assess their performance. These metrics are:
- Deployment Frequency
- Mean Lead Time for Changes
- Mean Time to Recover
- Change Failure Rate
- Monitoring & Observability
We worked tirelessly to bring in systems that automated the collection of these data points. The work we put out is great because of the high DORA scores. What we say now is verifiable. Hell Yeah!
Writing about our expertise has always been high on our agenda. This year, we went one step ahead. We wrote several new articles & tutorials. Some of these have even been cross-referenced in other publications and websites. We also started sharing our expertise and experience in video form.
We conducted several webinars with topics ranging from Observability, DataOps, AI, and more. Praveen from our team even spoke at React London. YAY! We aim to double down on this experiment and release more content in video form in the coming year.
We've also conducted several community events this year. Our sister concern, Tuesday, is now a Global Webflow Community Partner. We conducted webinars and in-person events at our Pune office to discuss the practices we follow and build a space for makers to share and network. We're super proud of this accomplishment and hope to continue building a thriving community of makers.
Our Playbook has also seen a few revisions. I was thrilled to see the team discuss the changes they wanted to see, edit the document, and publish changes.
As we go into the next year, we're going to focus heavily on three tenets:
- Verifiable delivery practices & metrics
Social and community are a given. We're going to use video & text-based content to share our experience & expertise.
Delivery practices are going to be a fascinating problem to solve. Our mission statement is to "Fight normalcy & transform a group of people into a team." We've always believed that small teams with the proper practices, skills, and context deliver exceptional results. We're going to continue to build on these outstanding practices. Take, for instance, a simple practice we follow called user testing. We get people from different departments to use the product we've built. The team watches as they use it; they spot where they get stuck or ask questions, then gather feedback to improve the product: such a small practice but excellent results.
It’s been over a year since ChatGPT was released, and as of date, it has 180+ million registered users. There is no denying that this is here to stay.
This new wave of automation will bring in very different problem statements and use cases.
One of the most interesting experiences I had this year with ChatGPT was when I introduced it to my wife. The first question I got asked was - Why do I need to sign up? and the next was - What do I ask it?
The conversational interface makes it personal, but using natural language can have side effects. The internet, at the end of the day, is a publishing platform. Billions of pages are published with a vast array of information on every subject under the sun. Making sense of this information and presenting it in a natural language interface makes you believe the answer it throws out is accurate. Unlike a Google search where you get a list of blue links and the rest of the web, a single paragraph as an answer (which may not be accurate) leads to a rather unique design problem. How do you show through a conversational interface how accurate or valid the information is?
What excites me, though, is the underlying LLM and its probabilistic model. The interface can change.
The use of this LLM, however, will require deep product thinking expertise, design and engineering excellence. Most companies are trying to fit it into their existing way of using their digital workflow, but we all know that over time, their digital strategy will change to fit the new norm. We don’t know what and how this would look for, let’s say, the digital strategy for an insurance company. Only time will tell.
While businesses adapt to this new layer of automation, there will also be a new wave of startups that are created to disrupt existing ways of doing work. So many billion-dollar businesses have been created out of just the options available in Excel. If Excel were an LLM, there would be vertical-specific solutions that entrepreneurs build and execute really well to create the next Quickbooks or SAP. Again, whether this is a thin wrapper or something else, only time will tell, but these are exciting times to be in.
Data is the new Oil/Sand/Oxygen
A lot has been written about Data. It’s been compared to being as valuable as oil and as abundant as sand or oxygen. The underlying truth, however, is that getting the right data will be paramount not just for businesses but also for governments to progress.
There is an abundance of data that is available out there, but creating a mechanism for it to be effectively available to make decisions is going to be a key challenge. This is a difficult policy decision and a bigger engineering problem.
Take, for example, the ambitious goal of going green. For a greener economy, data from industries, energy producers, transportation, and a lot of other sectors needs to be made available in a form that can be looked at for better decision-making.
We’re extremely excited about the possibilities with data in the coming years. Businesses wanting to increase profits and expand their moat will have to engineer their systems to collect the right data from disparate sources to find the path ahead.
2024 is going to be a roller coaster. We’ve narrowed our focus to just three areas of expertise:
- Data Engineering
- Generative & Applied AI
- Application Development & Design
We’ve also narrowed our service offerings to these four:
Each of these service offerings is designed like a product. We’re thinking of each service as having its own workflow where customers go through stages - awareness, signup, retention, engagement, and more.
I can’t wait for the year, and I’m looking forward to running more experiments and turning Wednesday into a brand and a preferred partner to ambitious entrepreneurs.