It’s interesting how we can learn and evolve by having great conversations. Conversations where we can share same or different opinions.
This was one of those cases…
Some weeks ago, during a coffee break I was talking with my great friend and work colleague Cornelius Engelbrecht. We started to share some past coaching experiences from previous companies, teams and the controversial topic of estimation.
This was the moment that we both agreed that we should share our thoughts and write about Startup vs Enterprise Estimations and why this can be “Two warriors with swords”.
“Our faces (Eduardo Ribeiro and Cornelius Engelbrecht) as soon we agreed to share our thoughts and write about this topic since is very controversial.”
As starting point, we would like to give some context and have everyone in the same page.
Nowadays there are two popular communities regarding estimation whom are perceived to have opposing views (The Swords):
On the one hand you have where its believed that estimation (guessing the size of a piece of work) can be useful and helpful with the following:
- A trigger for valuable conversations which help create a shared understanding within the team.
- To help the team to understand if the work needs to be broken down into smaller pieces to help reduce risk and complexity.
- If a team is using Scrum, then it can be a guide to taking on too much work in a sprint.
- To help us (Team and Business) in predicting/forecast when something will be done. Still, this topic is dangerous if not understood by everyone that an estimate is a guess/prediction and you cannot commit to a guess.
- It can help with prioritisation of work (e.g. deciding between something that should return high value and take little effort to do, and something that should return less value and will take more effort – its an obvious example, we’re sure you get the point).
On the other hand you have a more recent approach where people believe that we don’t need to estimate any work items, since the team has enough maturity to break down these items to a level where all of them are the same/small enough size (usually can be done within a week). Then you simply count the number of work items and then work on the most important items first. Using this approach you should be able to solve the same problems as in “View A”.
You might be asking… “What option works better and why?” or “Are they not trying to do very much the same thing?”
Well… when the #NoEstimation community started, their objective was to not use any estimation. For several years they shared their belief, and with time passing we started so see a change of opinion from no estimation to high level estimation using T-Shirt sizing on pieces of work that have not yet been sliced down to the smallest size.
What happened to have this approach change? Its unclear without more data or input from the community. By using high level T-Shirt sizing, are you still in the #NoEstimation camp, or are we are estimating? What do you think? ….. Or do you have a different experience with #NoEstimation? -We’d love to hear your experience.
Consider this: Since the beginning of ages we’ve used estimation to help make decisions and predictions. This is not an argument for estimating being right or wrong, we’re merely pointing out that its part of our nature.
For example: If you had an important meeting in another country and didn’t guess (estimate) how long it would take to get to the meeting (how long it would take to get to the airport, fly to the other country, then get from the airport to where the meeting is held).
How would you decide when you should leave, which flight you should take, what method of transport you would use to and from the airport, so you would arrive on time to the meeting?
We’re trying to point out (pun intended) estimation can be helpful and might make sense in some situations, even at work. Don’t you think?
Now it’s time to return to our main topic “Startup vs Enterprise Estimations” and why we think that are “Two warriors with swords!”.
Picture graciously provided by João Vale.
Warrior 1 – The Startup
Let’s look to a standard Startup where they have maybe a couple of teams and they own their product.
When we have this kind of start up at the beginning that have only one idea and they start developing their MVP.
So, they start from scratch (Idea) and they start developing where every iteration they deploy small value to their customers, gain feedback and incorporate the feedback in the next increment. Their objective is to evolve their product MVP to what their customers need (based on their feedback) and release as quickly and frequently as possible.
- When would a Startup find estimation valuable, why would they estimate:
When a Startup is not only driven by getting their idea out to their customers as soon as possible, they also are driven by need to know if its achievable before a date outside their control (e.g real world events such as SuperBowl or any other event or investors conditions). Therefore these Startups have to be more frugal on how and where they spend their money. To ensure they get the most value & return on what they spend and therefore can grow, estimating their work and including the value the work will return, will help them to frequently re-prioritise what to work on next – effectively shaping/changing their roadmap based on their customers needs. Think of it as a small boat that can quickly change their course/direction with minimal impact and reach their destination at the right time.
***As an example Tesla needed to get their first car (Roadster) out by a certain date, otherwise they lose their investors funding by losing confidence & trust in Tesla´s ability to deliver.
- When would a Startup find no estimates valuable, why would they use no estimates:
When a Startup main drive is to get their idea built and into their customer´s hands as soon as possible, it does not depend on a specific date… just as quickly as possible, then using the #NoEstimates way will make sense. This is where you slice the work into the smallest possible pieces (fitting into a week or less), incrementally and iteratively working on the most important & valuable piece, and just count the pieces of work to get an idea of where you’re going. This is still the small boat that reaches that can quickly change their course/direction with minimal impact and reach their destination as quickly as possible.
***As an example when Elon Musk created SpaceX, his vision was Mars colonisation. Of course he would want this as soon as possible – and if it could be by tomorrow then it would be great, however it is not need to be by a certain date, just as soon as technically possible.
Warrior 2 – The Enterprise
Now it’s time to look to a bigger “Beast” aka the typical Enterprise where we have many teams, departments, locations, cultures and have several products being developed simultaneously whilst keeping investors trust in confidence in our ability to deliver.
- When would a Enterprise find estimation valuable, why would they estimate:
Most Enterprises have the added complexity of co-ordinating multiple teams delivering multiple products across multiple locations and multiple platforms (… you get the idea) whilst being subject to Marketing campaigns that are prepared for release on specific dates that can be Online, Newspaper, TV, Radio, etc… and which are connected to real world events such as a Superbowl (once more… you get the idea of complexity).
Estimation help create a shared understanding in the teams regarding the work, promotes conversations which leads to the discovery, and help make decisions on what to do due to the cost/value and return.
Estimation can also be helpful for Enterprises to forecast when something can be done, in guiding decisions when co-ordinating teams/deliveries/products and their inherent dependencies whilst adhering to governance and regulations. They would still break down work into smaller pieces, and prioritised the most important/valuable work first,however this priority/order is influenced by synchronisation of many parts of the system(teams/products/needs). By estimating and knowing the effort involved you can take smart decisions and plan your roadmap and the co-ordination required by the many different parts of the system to deliver by a forecasted date.
This is not a small boat that can quickly change their course/direction to reach their destination (think oil tanker…).
- When would a Enterprise find #NoEstimates valuable, why would they use no estimates:
In theory when an Enterprise could use the #NoEstimates way when they have completely independent products with cross functional teams/departments working on goals that are independent and not driven by external markets and/or events. Unfortunately we don´t have much data or any examples of where this works on Enterprise organisations – If you know of any please feel free to share this with us/everyone, we’d love to learn from you!
For our final thoughts:
If you pay attention you will notice there is a small battle happening between #NoEstimates and Estimates, and these are the two swords each warrior has, and the Startups and Enterprises are the warriors.
We hope you realise these are two different realities and we shouldn’t blindly support the one camp or the other and apply the same approach to a Startup and to a Enterprise.
Keep in mind that this can create a higher risk and create bigger issues rather than fixing it. These approaches are not Silver Bullet´s to solve all the issues you are experiencing.
So, before taking any decisions is important to understand both sides and not just join this battle with your decision made.
One needs to have a good understanding on #NoEstimates and Estimates in different business realities and what benefits or impacts they can bring.
Ask yourself why we estimate? What problems are we trying to solve with the estimates? Are we trying to solve the right problems?…..
I’d say it boils down to a specific question: How would you make a decision without an idea about the size of the impacts of that decision?