Sunday, March 9, 2014

Decentralizing product management for shorter cycle times

For any enterprise software company or startup, the ultimate measure of success is how much money it makes for its investors. Money is the best surrogate we have for value a company delivers to the world - and by extension, the impact it has on the world. 

For a product company like ours, its the product that delivers this value. Hence, the goal of product management is to build products & features that customers value & are willing to pay for. Our customers derive value when they use our products to address important design problems, but the journey starts much earlier, starting with how they discover our product, their perception of the product pricing, their experience during the sales process, and most importantly - their experience after they have bought the product.

On the other side, in order to be successful, product management works with designers & software developers in defining and sequencing what needs to be built to deliver value.

Development cycles are no longer a problem area

While the latter has received significant attention, and almost down to a science. Thanks in part to the rich literature on development, over the last year we have continued to drive down our time-to-production, which today stands at 5 days (now with much reduced variability since we moved away from Scrum & adopted Kanban). This means that any new feature that customers would value can be built & deployed in less than 5 days. A major contributor here has also been our focus on moving developers closer to customers via support rotations, frequent usability testing, and integration with engagement-tracking software like Totango which give us rich visibility into how the product is being used. We don't ship any new feature without tracking commensurate module/action in Totango.

Time to revenue as a metric

While for software development teams Time to production metric an objective works well, one of the metrics to track for product management becomes Time to Revenue, and minimizing this time requires significant focus in ensuring that information flows from the edges of the business to product management, and the other way around, as quickly as possible.

Now unlike software developers and designers, their primary job is not to respond to product management's needs. Many constituents of the business interact with the world, and their primary job ranges from marketing & selling the product to helping customers adopt the product in greater numbers, thereby making money for the business. This makes flow of information inherently frictional, and the business must organize to make this as frictionless as possible

In order to solve this problem, most companies have product managers who go & speak with customers, sales teams, marketing teams, and customer success teams to learn about the world outside, and in turn helps them do their job better by arming them with valuable information. This makes product managers an important constituent. This also implies that there is reduced burden on other business teams in deeply understanding the needs of product management.

Decentralizing product management

At Sefaira, we have taken a different approach. We believe that organizing our teams to maximize our surface area that interacts with the world, product management needs to be distributed at the edges, and product management responsibilities need to be embedded with the edges of the business. In order to be successful here, we need to ensure the teams are always aware of key things we are trying to learn from the world, and in turn be able to push this information to product management. Conversely, they need to be the voice of product management. 

Product management responsibilities need to be embedded in the DNA of all business functions. This reduces product management's responsibility to organizing incoming incoming information, look for the most compelling hypothesis and test them by devoting maximum development fire power at these hypothesis. In turn, product management leans on the business to evaluate the outcomes.

What does this mean for other functions?

For different stakeholders, I will outline key areas where they need to think like a product manager in addition to their core responsibilities.

Sales rep

  • When a prospect ask about functionality that the product doesn't have, learn if the prospect would value the functionality they are seeking. If yes, what problem would they solve with it?
  • When a prospect is unwilling to buy at proposed price, learn how much they are willing to pay for the product
  • When a prospect is unwilling to buy, learn what problems the product would need to solve for them for them to value it
In turn, they should be equipped to talk about the kinds of problems we are looking to solve in the upcoming months, and learn if prospects value solutions to those problems. A false or a positive signal both unlock different kinds of investments based on product management's judgment.


  • Build new content and drive awareness based on what the product team is building
  • Create and spread the word on success stories
  • Build a compelling set of use cases for the product 

Customer success

  • Learn of blockers to product adoption
  • Recruit users for usability testing and user interviews
  • Learn how latest features are being used & provide insight to product management
Once these expectations are clear, the organization needs to be instrumented such that information can flow freely. We intend to rely on a combination of Asana, Slack, & Totango to stay aligned here.

The benefits of such organization are immense. Due to inherent decentralization of product management, this can shorten cycle times significantly, and scale well as the business & product footprint grows without the need for hiring additional product managers, or impacting cycle times. This keeps our business responsive to customer needs and allows us to bridge the gap between problems and solutions rapidly.

Monday, February 3, 2014

Amazon's Last Mile Problem (Hint - It's not Amazon)

Amazon recently announced that it is considering a significant increase in annual subscription price for Amazon Prime. Amazon Prime, for those who live under a rock, gives its subscribers free 2-day shipping on almost all Amazon purchases. This price increase will net Amazon significant revenue. However, it is fair to assume that this will increase pressure on Amazon to deliver on its promise of 2 day deliver - something observers claim doesn't always happen today.

This brings me to what I see as Amazon's major challenge - the last mile of their delivery chain which is when the goods meet the shopper. Of the 2 occasions I've been disappointed by Amazon wasn't Amazon's fault.

On the first occasion, I wanted a package rerouted after it had made its way to the local fulfillment center. I called Amazon's customer service requesting a re-routing of my package. However, after 10-15 minutes of phone conversation with Amazon (in which time they basically told me nothing more than what I looked up by tracking my FedEx package), they simply referred me to Fedex saying they are unable to change delivery. I called FedEx, and as I expected, FedEx simply said my request was simply pointless.

On the second occassion, I ordered a pair of running shoes. I put 25-30 miles on my shoes every week, and few things give me as much joy. I was looking forward to getting my hands on Mizuno's Wave Rider 16 (that is one good looking pair of shoes!) like a little kid waiting for Christmas. I ordered them to be delivered at my office since it pretty much guaranteed that I would be there to receive it. I waited till 8:30p on Friday, only to receive a notification from FedEx that no one was available to receive the package. I was thoroughly disappointed.

What is interesting about the two events was that Amazon wasn't to blame, it was FedEx. Yet, all my disappointment was directed towards Amazon.

As a shopper, I want my goods faster & cheaper than ever before. Jeff Bezos gets it. I am certain that Bezos gets the sheer joy shoppers get from getting their hands on things they want, and I am certain he doesn't like being vulnerable to courier delivery services that culturally not as in tune with the customer as Amazon is, nor are they as motivated to improve as Amazon is.

How Amazon addresses these challenges will be interesting to watch, especially in light on increased subscription for Prime. I will continue to pay... for now.

Monday, January 20, 2014

Importance of free will for employee engagement

Like many other companies, at Sefaira we realize that empowering and engaging our entire team with the company's mission is critical for our success. This implies those closest to the front lines, i.e. our customers and our code, should feel complete engagement with the goal of the company & motivated to make decisions and act in the interest of Sefaira's success.

The business value of employee engagement isn't suspect, and can be proven by empirical data. Anecdotally, we feel this helps with improve employee happiness as well. However, I recently read a study by Mihaly Csikszentmihalyi‎ which provides empirical data in favor of this.

Using his novel Experience Sampling Method, Csikszentmihalyi‎ studied a population of workers across a wide spectrum to understand when people enjoyed themselves. He learnt the following -

  • While at work, 54% of the population said they experience enjoyment. During this state of enjoyment they felt challenged & they felt their skills were being used well
  • While at home or during leisure, only 18% of the population said they experience enjoyment. Interestingly, a majority of workers felt apathy & boredom during leisure hours
This clearly showed that people enjoy themselves more at work than they do during leisure hours. However, a 2nd question showed an interesting paradox. A majority of workers who said they were enjoying themselves at work said they'd rather be indulging in a leisure activity. Also, a majority of workers who were bored during leisure stated that they'd rather not be doing something else. 

This means that a majority of workers would rather be bored & apathetic, than engaged in enjoyable work. This was a paradoxical finding. The author states that this can likely be attributed to the social perception that the nature of work is not determined by employees (even though work itself is challenging & enjoyable), and is not under the employee's free will. This lack of free will leads employees to choose to spend their time being apathetic, away from work they truly enjoy.

For us as managers, the lesson is clear. The more employees feel like they are acting under free will, the happier they will be, as they will seek the enjoyment that comes from work. Of course, the business benefits from this if it is organized to allow employees to decide & act in the interest of business.

Tuesday, January 14, 2014

Two ways in which quotas can debilitate startups

Quotas(1) play a very important role in businesses. They are the chief instrument for translating business objectives into measurable milestones. Further, they translate measurable business goals into objectives for revenue generating functions in the business. Ultimately, they connect all other functions ranging from product development to marketing with overarching business objectives.

However, they suffer from two major weaknesses especially for startups who tie themselves to quotas like established businesses would.

Quotas are not actionable metrics

I was first introduced to the idea of actionable metrics while reading "Out of the Crisis" by Edwards Deming who used accident statistics as an example of an outcome-based metric that measures the result. However, it fails to provide any insight into what actions can be taken to improve the metric. Similarly, quotas are an outcome-based metric, which can't translate into actions as highlighted by Jason in this HBR article. Naturally, as Tom Tunguz points out in his follow up, they lead to anxiety in absence of actionable metrics (which are often process indicators) as they leave sales people powerless (2). Actionable metrics are important for sales as they result in constancy of purpose, and break business objectives down into fundamental metrics that can be acted on. In essence, actionable metrics translate sales into a machine that looks for continuous improvement as opposed to attainment.

Quotas can obfuscate reflection

Established businesses have a business model that works, they have a sense of market size and their customers, and therefore - business goals well informed. Startups are unlike established businesses. Business goals aren't informed by an objective market size, as very often startups are creating new markets based on their offerings, or fulfilling needs that the market didn't realize it had. This implies that unlike established businesses, startups are going through constant learning. A natural extension here is that startups  need to reflect about their sales processes independent of attaining quotas - way more than established businesses do. The challenge here is that quota attainment (or meeting budget) can sometimes result in inadequate reflection, and delay in actions which can debilitate a startup.

A question for more informed folks, especially those that sit on startup boards - Does measurement of leading indicators lead to better "retrospectives" by executive teams than meeting budget?

1 - I use quotas interchangeable with budgets here. I find that the difference between the two is essentially their probability weights. Budgets have 90% likelihood of being met, where as quotas might have 50%.

2 - In the face of goals, being powerless causes anxiety - a thesis proven by Mihaly Csikszentmihalyi in Flow