Not All Scalability is Created Equal II

Not All Scalability is Created Equal II

After last week's writing, this section hopes to lay out a few more specifics about trade-offs companies make that enhance or diminish their scalability. There are some specific examples and references to LearningLeaders below as well to put it more into context. Next week, I'll wrap up with a more thorough analysis of what scalability means for the education industry.

So what does it actually mean to be scalable? What are signs a company is scalable?

Reverse engineering scalability and its ingredients requires us to work backwards from the lagging indicator: profitability. Oh boy - so this means that essentially anything that goes into keeping profit margins higher (or at least flat) as the company grows can be a factor in whether or not the company is scalable. While that’s technically true, there are some factors that from my vantage point are more relevant than others. Some will be discussed below here and many more will go unnamed in this writing because they are negligible, they are not occurring to me at this moment, or I am simply ignorant to them. (Note all of these depend largely on industry and geography as well!). Some of these factors may be root causes of generating scalable growth, while others may actually be the products of scalable growth in a first stage and then causes of ongoing scalable growth

(ie. Network effects). The edges between leading and lagging indicators are sometimes blurred and candidly I need to think more about them - but here’s a first draft:

1. Clarity of Market Focus and Customer Awareness: Focusing on Ideal Customers vs. Reactive to Customer Selection

Actively focusing on specific segments of the market is an imperative component of scalability because it forces us to refine the offering itself and the unique value proposition offered into something more focused. The least scalable situation is one in which the company is completely reactive to customers and will sell to anyone who walks through the door. This may sound counterintuitive because scalable companies are mostly large, right? At times there are overlap between the two, but not often. The largest companies in the world and the most profitable have a clear idea of who their ideal customers are and create products, services, and experiences for those key personas. Apple, Samsung, Royal Dutch Shell, etc. all have a very clear idea of who their customers are and you better believe they are NOT reactive in their customer selection. They hunt their ideal customers incredibly well.

Not having an ideal customer in mind means that you are wasting marketing dollars chasing potential customers who won’t buy. This also means you are accepting customers who won’t appreciate your product and who will cause challenges later by way of complaints or refunds. While unavoidable to an extent (there's been no company I can think of with 100% satisfaction rate) the more customers you take on who aren't vibing with your company or product selection cause margins to suffer due to the above-mentioned wasted marketing dollars. This also means you will be unable to streamline your customer acquisition process and thus the sales process will become more expensive over time.

By clearly segmenting ideal customers (which by the way, can still be millions or billions of people - this number does not need to be small), a company is able lay out the foundation on which to build an enterprise with highly scalable margins.

Let's define and focus on ideal customers.

2. Customer Acquisition Focus: Customer-Centric vs. Sales-Centric

A Sales-Centric process is designed around providing explicit actions to salespeople to encourage them to sell the products, services, or experiences on offer. This process (of course better than having no process at all) may be more expensive to scale because it depends on specific people’s talents - typically ’star’ salespeople, it is more expensive because it actually requires that salesperson’s time, and it’s possible for salespeople to take shortcuts in the process to get to the finish line rather than address all the concerns of the potential customer. Products sold in this way are more often commodities, transactional, and volume-driven. Salespeople are 'agents,' rather than team members. In this setup, reaching large sales volumes is not the issue, the challenge is maintaining the margins around the sales process unless you have very high-ticket items.

On the other hand, a Customer-Centric process is designed around ‘helping the customer buy.’ Each phase of the process is designed to be cognizant of the emotions and decisions each potential customer will feel and make at a specific step within the buying journey. The customer acquisition process is then designed to help accelerate and facilitate those decisions. The reason why this method is more scalable over the long run is that it’s less dependent on ’star’ salespeople, the process educates the customer throughout the buying process so there are fewer post-sales problems, and the process encourages referral business (the most scalable form of customer acquisition out there!)

By increasing time and efforts to ***eventually reduce the cost of time per RMB of revenue (notice: not sale)***, we can effectively keep customers buying without edging into net operating margins. This may require more time up front, yes! But if we spend 1 hour with a customer who then refers two more, rather than 30 minutes with each, it's an easy trade-off to make! If you haven't guessed by now, this is also why retention is so much better for our company than new sales - the amount of time for each retained customer is a fraction of new customer.

Let's have a customer-centric sales model.

3. Offering: Outcome Focused vs. Feature Focused

Customers care about outcomes. That’s because outcomes are what actually affects them! The majority of customers don’t really care about how much work went into creating the product, service, or experience. Some might, but the majority of customers don’t really care about the bells and whistles or the features themselves. They care about what the product or the experience does for them. They generally care about the end, not the means.

The longer we spend tooting our own horn about what the features, advantages, or benefits are of our work with the customer, the less they will care. Limiting risk (variability) and increasing certainty of outcomes the customer desires should be the goal of communicating the value of a product, service, or experience. For the debaters among us, we want to: increase magnitude of outcome, decrease risk that it won't occur, and push forward the timeframe so that it happens as soon and as often as possible.

In other words, show the customer that 'Ah-Ha Moment' as early and frequently as possible to ensure the desired results are being delivered.

Why does focusing on outcomes increase scalability? Customers know the result they should expect from using your product, service, or experience. They don’t need to put the puzzle pieces together themselves

There’s a simple ‘Yes or No’ calculation in their head as to whether the offering solves their problem. At that point, the only remaining decision is price: Is the amount of investment in this product/service worth the outcome on the timeframe provided: 'Yes or No.' This all increases pace of customer acquisition.

Let's deliver powerful outcomes early and often.

4. Technology and Efficiency: Repeatability vs. Random

Companies that deliver high levels of scalability are decidedly repeatable in their nature. This has existed since iron smelting and railroads and even before that, so it's not a twenty-first century invention. But I do believe it is a twenty-first century phenomenon. While I cannot be 100% sure, it's my estimating that the reason why so the word 'scalability' is popping up so much in the last one to two decades is largely because of Silicon Vally-based technology businesses. 'Copying' a product or service digitally (repeatability enabled by technology) for near-zero marginal cost has transformed the business world forever in innumerable ways. A key enabler of scalability in businesses is the copying and moving of digital assets. If a business is able to do this successfully and sustainably, then they will have an advantage over rivals who cannot. The repeatability of this process have driven many companies to digitize their offerings over the past decades. The trend won't stop, but accelerate. Digitization allows for the repeatability of any offering to sky-rocket. It also moves costs from a variable cost profile to a fixed cost profile - this is going to be important for the education industry specifically as I'll write about in the next segment.

By driving as much of the daily operations into instantly repeatable processes, we improve our throughput, improve quality standards, lower cognitive load on our team, and move more costs from variable to fixed. It's a huge transition process but what sustainably scalable organizations have been successfully navigating for quite some time.

5. Company Operations: Parallel vs. Sequential

As a company, operating in parallel means that multiple versions of the same process can happen at the same time. Sequential means that one process necessarily needs to happen after the other. When a company is able to operate in parallel, it means there are multiple people who can perform the same function and multiple functions can happen simultaneously. That is great for scalability. When a company is able to operate in parallel, it also means there is very likely to be systems in place to drive those tasks to completion. Operating sequentially doesn’t imply that, as the same person or team can perform the same task repeatedly without needing to create or record a process for it. This also creates a culture of 'stars' where processes grind to a halt if a certain person doesn't execute on their plan.

The parallel processes can be in existence at once, the better for the businesses ability to preserve margins as it grows. Fixed costs once again are distributed among all of the actors in the value chain.

Working in parallel means planning in advance so that work streams are able to operate without competing for resources or waiting for feeding buffers.

6. Culture + Trust: Building Trusting vs. Losing Trust

The more people within our company trust each other, the faster we can move. Businesses only move at the speed of trust. This is because there are so many implicit and contextual components of decisions that surround us every day in the decisions we make. In my opinion, the best companies find ways to create trust among the team. Google's Aristotle project even proved that the number one factor shared among successful teams is psychological safety - team members feel safe taking risks and being vulnerable in front of each other. Perhaps controversially, I believe this is best done through a strong culture, because team members are more likely to know what to expect and how people will react. Remember that culture is not necessarily good or bad, but it can be strong or weak. ‘Competitiveness’ or ‘Results-driven’ can be great elements of a culture, though not at every company. The same can be said about ’Teamwork’ or ‘Compassion.’

When a company decides on its core values and the team actually lives out the core values, trust is created. At every marginal decision-making opportunity, if the core values (or virtues, leadership values, driving ideology, whatever you’d like to call it) are present and strong, decisions can be made faster without approval from the big boss because less-experienced or more junior team members understand and know what the senior management might want them approach the given situation.

Trust is never static, but always increasing or decreasing. It is up to the senior leaders or management or more experienced team members who the company looks to for leadership to take the first step to create a psychologically safe place to make suggestions, dissent, and propose new wacky ideas. Of course, not every suggestion or request can be implemented, but if the team is on a path to continually building trust, then decisions can be made faster and faster.

In a company where people talk behind each others' backs, spread rumors, or gossip, this is precisely what destroys trust. Team members are less likely to speak up because no one knows who is going to re-purpose those comments for their own use later on. This kills trust, speed, and scalability within a company.

Why does a company with a strong culture and high level of trust create more scalability? Because cost of company operations drops - you can work in parallel instead of sequentially. Because cost of management systems drops - there need to be fewer ‘check-ins’ and ‘checkpoints’ along the way. The faster and more autonomously the team can operate, the fewer additional costs that are imposed on the system preventing margins from being maintained or expanded.

Let's prioritize being open-hearted as leaders over always being right. It builds trust and allows us to move faster.

So if scalability is so great, then why doesn’t everyone do it? What are the challenges?

If creating a company that grows and maintains or expands margins was easy, then everyone would do it. But of all the businesses that are started in the USA every year, it's estimated that only 4% reach the 1MM USD Annual Revenue Mark and only 0.4% make it to 10MM USD Annual Revenue. There are some industries in which scalability is simpler (not necessarily easier, but simpler). Software is the most commonly-used example. For every Microsoft Office 365 client that exists, Microsoft pushes a similar application to everyone’s browser.

(Tech team at this point is shaking their heads saying, “Yes, but you need to load-balance and ensure that processor speed isn’t sacrificed when multiple users are operating simultaneously...” That’s why I said simpler, not easier. Every form of scaling has its own challenges, but software at least does have easier replicability of product!)

What are the major challenges while scaling? There are many and not every one is able to be solved at every stage of the development of the company. When the number of users doubles, perhaps the company is really excellent at scaling one function but not another. Perhaps marketing does a great job reacting to doubling of inbound leads, but the product team isn’t quite prepared to handle it. Or vice versa. I know from firsthand experience at LearningLeaders we've faced a number of these each and every one of the challenges below.

A few of the largest challenges that prevent companies from scaling, to wit, are: 1) Variation in Output, 2) Efficiency of Agents, 3) Complexity of Processes, 4) Comprehension of Entire Systems and Fragmentation of Knowledge, 6) Morale and Feelings of Powerlessness.

This is hard: Plans to Reduce Variation of Output

The problem of variation of output is the single-most critical preventative factor from a company actually achieving scalability (And this is the one that online/digital products have less often, as mentioned above). If McDonalds cannot produce all of the burgers to taste the same, then they will violate their brand promise and will not see repeat customers. If LearningLeaders gives a wildly different experience to every single student who comes through our doors, then parents and students won’t know what they are getting themselves into. Thus, the risk of working with LearningLeaders increases and the perceived value for a certain segment of customers drops. And, by the way, the variation of output is only set to increase as n, the number of units produced/served, increases. When we had 100 students it was far easier to control the output of each course/class than today. This is exactly why GE was famous for implementing Six-Sigma methodology to ensure quality control and that all output quality was similar - the quantities of goods they produced became so large they needed an entire system and division just to measure the replicability of output!

The way to combat this challenge of variation of output is to standardize processes of inputs and development, leading to more standardized output. Oof! I’m sure that last sentence rang of blasphemy to any educator reading this. Though, I hope once we examine it further, that doesn’t remain so prickly. The purpose of standardization is not necessarily to curtail a certain student’s creativity or personality. It’s not to clip their wings. Rather, as many before me have said, standardization's goal in education is to provide the baseline skill set and capabilities to everyone. Then we can allow each student to specialize or focus in whatever area they wish once the foundational skills have been mastered. This is a tricky balancing act and one that I contend does not have one perfect solution. What's worked in China likely won't work in Chile. What's worked in Finland likely won't work in the USA. There are many brilliant people who have tried to delineate this boundary over the last hundred years or so (and even longer) in education systems all across the world and no one can agree on exactly where or how to draw the line. If anyone would like to read more on this topic, I highly recommend Education and the Commercial Mindset, by Samuel E. Abrams.

My stance here is that consistency matters most. No matter where the line is drawn (notwithstanding all of the natural disagreements that will arise as to where it in fact is drawn) all actors and contributors within the system must be consistent in their actions supporting students towards these prescribed levels of competence and mastery. Moving forward, with our improved curriculum debuting in the Fall of 2020, I'm looking forward to an even more systematic experience for all of our students.

This is also hard: Efficiency of Actors Carrying Out the Plans

Another challenge with reaching any form of scalability in a human judgement-driven business is that the more actors there are executing a certain plan, the greater the level of difficulty it is to have them follow the procedure with similar levels of quality. This is not at all because any of the actors want to malign the process but simply because each of the actors approaches the process or plan with their own biases and tendencies, just like any normal human being!

I’ve made these mistakes many times myself when in the classroom with students, so I know first-hand how difficult it is. For example, twice or thrice a semester, I forgot to ask students for their TILT at the end of class. This is an absolutely critical part of the procedure of operating a class as LearningLeaders - one that is scientifically proven to prime students’ brains to retain more information and experiences, one that differentiates us from other programs, and one that readies the students with a more confident answer the classic question from Mom or Dad at dinner that night, “What did you learn at debate class today?”

But since none of us are robots or computer software, it is extremely challenging for us to perform the procedures perfectly every single time. Even if there is a procedure or lesson plans, as discussed above, there's no guarantee the actor will follow-through. Unfortunately for many educators, actually following a centralized lesson plan feels like very definition of stifling their creativity. Frankly, it's true! In fact, by very definition, the following of a procedure will likely restrict creativity! But if consistently great outcomes are the desired result, then following the designated procedures (essentially checklists) have been proven to increase both quality and consistency. A great book on this topic is The Checklist Manifesto by Atul Gawande. Doctors, airline pilots, and career all benefit tremendously from checklists.

My viewpoint is that we all must sacrifice some of our creativity or personal preferences if we want to work with others on a mission larger than ourselves. I have been impressed over and over about how LL coaches are willing to swallow their pride and do what the students and parents and organization asks and expects them to do - even if the activity or final sentence in a lesson plan doesn't perfectly match their wording. In order to coordinate on a larger mission than our own lives, finding some sort of compromise is necessary. Some people cannot handle that type of restraint and so they work as solopreneurs or as individual contributors within larger organizations. Many impressive artists simply don't want to follow the procedures outlined and so decide work solo. There's nothing wrong with that.

Hey - even Beyonce left Destiny’s Child to pursue her solo career. It's worked out pretty well for her. Good move, Bey.

This is super hard: Complexity of Process

The larger the system gets, the more likely it is that variables in one place affect results in another. This leads to complexity. The system moves from simple relationships - clear linkages between independent and dependent variables - to a more complex system of multi-variate interdependence. The more variables that exist within a system, the more challenging it is to determine with any sort of conviction the predictive capacity of changes to the independent variables. This gets to the very heart of the Education Crisis that's happening in my home country of the U.S. right now - there is a raging debate as to whether a greater impact on the testing results in many school districts are from the teachers, the curricula, or the families and households from which the students come.

Within each of those three larger segments there are dozens of individual variables and the complexity soon becomes more than any human (and currently any computer) can solve. Imagine how frighteningly complex the systems of companies like ByteDance or Uniqlo or Cadbury are compared to LearningLeaders. And yet, they are all pursuing business plans with high levels of scalability relative to their industry. Over time, organizations (and organisms, actually) become more and more complex.

At the start, the amount of variables we needed to consider for what impacted student success was not few, but certainly not as many as we do now with larger numbers. We were able to easily speak face to face with 100% of parents and students, and understand where LL was meeting or surpassing their expectations. More importantly, we were able to quickly find out and isolate where LL was failing our customers, both students and parents, and adjust accordingly. Further, since we offered fewer courses to a tighter age range of students and from fewer schools, all of our thinking and attention wasn't spread as thinly across multiple offerings, but we were able to focus intently on the fewer variables that we had to consider.

My thinking is rather simple and perhaps even pessimistic here, I’m afraid: we must be content in not knowing. Understand that you won’t be able to perfectly comprehend the relationships of all variables in the system, but be content in finding the ones with the most statistical leverage (Side note: I think more students should take statistics, rather than calculus. In my experience, life is more about dealing with uncertainty than by actually solving for an answer. Different topic though). Trying to over-optimize every single variable will be a full-time job and then your work may be rendered obsolete when new variables are introduced. Rather, focus on a few variables that the greatest predictors of the desired goal. Unsurprisingly to you, dear reader, the single variable I believe that we should seek to maximize for ongoing company success is student retention rate.

Whoa! Really hard: Comprehension of Entire Systems and Fragmentation of Knowledge

Another problem with systems as they grow larger and more complex is the less detail human decision-makers can maintain over the entire organism. This often leads to making decisions to support a local maximum, rather than a global maximum. I’ve written about this before, but we can briefly examine here why this happens and what can be done about it.

Looking at this from LL's own past, at the start it was quite simple to recommend to parents a certain course. Now, since we have various courses, camps, and competitions students go to, there are greater challenges with understanding how they all fit together. A Community Team member at LL now is expected to understand all the courses, camps, and competitions that we offer when speaking with parents - that is incredibly challenging to complete with any degree of nuance! A People Team member working on recruitment ideally should understand exactly what characteristics coaches who work on every level of the program have so prioritizing recruitment targets becomes clear - that is remarkably difficult to do!

We've currently tried to tackle this issue through both inventing some of our own methods and adopting some best practices from other organizations: 1) Knowledge-sharing platforms (we have the KnowledgeBase), 2) Management updates (we have the Monday Memo, the Monthly Exec Team Memo, and the weekly Experience Team Memo), 3) All Hands Meetings (we have ours on Wednesdays), 4) Incentives (we do our best to align personal goals in Key Results Areas to company-wide priorities), and 5) Onboarding practices (we include quite detailed information about the company in ours compared to most other companies). We’ve also tried other initiatives to address this issue, including: Mike’s Office Hours once per month in Pudong and Puxi and responding to questions from the Partnership at All Hands meetings, among others. All of these communication platforms serve to provide context and detailed updates around what one part of the organization is doing.

The trouble with comprehension of the entire system is that at some point it actually becomes counterproductive. The level of detail that someone in Team A has over Team B’s work doesn’t reach the level of understanding that Team B has anyways, yet someone from Team A wants to provide their input to Team B and then gets upset when it’s not taken into account. Imagine a company with five or seven different divisions or product lines. It simply doesn’t make sense for all of the team members in Team A to be as knowledgeable about Team B's work as Team B is. Of course, none of us want to meddle in other people's work. And it also feels good to be 'in-the-know,' and comprehending what other teams are doing.

Soon, though, we'll need to begin to let go more decisively of feeling the need to understand everyone else's work streams.

My take on this: if everyone has their head in the clouds with their feet on the ground, we’ll do amazingly well. What do I mean by this? Head in the clouds means being able to understand the overall vision, mission, and values of the organization, the high-level strategy of how we are going to get there, and the trade-offs we are willing to make as a company to reach our goals. Feet on the ground means focusing on the contributions that this specific person can make to achieving that end. From the start of LL we've sought to have at least centralized knowledge-management and documentation of knowledge platforms as well as training and leadership development programs. I feel these should continue, but product-specific or event-specific trainings or information will only need to be shared with the relevant people moving forward. This is a challenging transition we are going through now where we are used to be aware of every change that’s happening in the company and now we are becoming more and more distant from changes as they’re made. There's no doubt there is a feeling of discomfort, which may lead to the final challenge preventing organizations maximizing scalability.

Wildly Hard: Morale and Powerlessness

The problem here is quite simple - when most people feel powerless over a decision that is made that impacts their life, they can feel demoralized and unimportant. This of course depends on the social contract (culture) that is established at the get-go, but also what decisions contributors feel that they want to have an influence on. Within a population of n size, x amount of people really want to have an influence on decision a. y amount of people really want to have an influence on decision b. And so on. As the number n increases, the probability decreases not only that all n people get to realize the decision they want, but also that sub-groups x and y get the results they want in decisions a and b.

At LL we've seen this multiple times, unfortunately. Someone really wanted x, but the result was y. The more people involved with and impacted by the decision, the more people will be disgruntled. But that won't stop us from hearing insights from the team on issues that are pertinent to our ongoing success. There isn't a way for us to implement every idea, but I do feel we can make ourselves as open to explaining why any decision that was made was made. It's unlikely everyone will agree (and, in fact, decisions that everyone agrees with are probably the more straight-forward decisions anyways...) which regrettably doesn't do much to help morale or make people feel more powerful.

My take on this: keeping teams generally smaller gives more people the type of influence and autonomy they desire. The now-famous Amazon 'two-pizza team' rule is a great rule of thumb: don't have teams bigger than can eat two pizzas in a sitting (typically 4-8). If the team is bigger, then split it into a separate group with more focused objectives. They'll get in each other's way less and each person will have a larger impact on the decisions made within the new team. Increase autonomy at all levels of the company by keeping teams leaner.