Look beyond numbers to achieve challenging goals

Everybody loves to achieve challenging goals, right? I love it! Gives me a kick. I have been evolving on how I approach my goals. Back in my school, I was more of a doer (without much planning); in my college, I learned how to set goals and solve problems by breaking it into smaller steps. During my corporate stint, I learned number driven problem solving where each part of the solution has its own objective metrics and targets. Being number driven not only helped me in removing my bias in creating a better solution, but it also helped me in aligning my peers and team members to my solution. I do enjoy having these tools in my arsenal to solve problems better and achieve challenging goals. But what’s next?

What's next to achieve challenging goals?

As I interact with many fellow entrepreneurs, I have seen them vent their frustrations with the unexpected negative consequence of relying on numbers to solve problems. I have seen two major issues -

1. Limited numbers & measurements - Many a time we struggle to create data required to solve problems. This leads to either non-optimal problem solving or significant investment in terms of money and time to create more numbers. In the latter case, decisions get delayed and companies bleed. For example, if we do not measure our key numbers, it becomes impossible to create focused direction for team members
2. Too many numbers - A complex web of too much data (aka data pukes) makes it difficult for people to align with a broad picture. For example, if you give 10 different performance evaluation criteria to a person and make 10 subcomponents for each criterion (as per company needs), the person will flunk in all

In both the scenarios, convincing peers with your solutions, becomes even more difficult and all the reliance on numbers end up in analysis-paralysis. 

Turns out that such problems are very common and I failed to find any relevant article (I Googled!). I will talk about my suggestion on where the problem lies and how it can be solved.

Where reliance on just numbers lead to?

Let’s consider a very generic entity that we want to optimize over time and rely on numbers for the same. It can be the performance of a team member or scorecard of a child or even P&L of a company. Or it can be ever more generic entities such as eggs we eat for health or pills we take for easing headaches.

I have used an iceberg to denote such entities in my blog (you will soon realize why). All these entities have various attributes,  for e.g in the case of an iceberg, some of them could be easily measurable, such as height, width, average speed, etc. Other attributes may be difficult to measure for e.g. weight or any aesthetic attributes such as picturesque, etc.

When optimizing these entities, we put objective metrics and targets on its various attributes. Initially, I have seen people struggling because we haven’t defined measurements and so it becomes difficult to align everyone and move forward. Hence the priority becomes to invest in terms of money and time to define attributes. After a certain point, we have a lot of numbers and measurements but by then the system becomes impractical for most of the people who need to use the very system. Reason being that human minds only absorb simple things for most of the mundane routines.

Important to strike the right balance

Even the most efficient systems would stop investing in creating more numbers just before it becomes “too many numbers”. This balance is important and it is important to understand that you will have “limited numbers” to define complex entities. Due to the limitation of measuring various attributes, we can never define an entity completely with just numbers.

Usually, numbers cover a very small proportion of entity. Most of it remains invisible from the numbers perspective. As we rely on numbers to communicate everything, we lose visibility of a big part of the entity. Yet we need to define the entities and most of us rely only on numbers for the same. Objectivity acts as an aphrodisiac to an analytical mind! Let’s see where it leads to?

What follows next is the usual optimization cycle based on this limited visibility. We optimize how our entities are perceived & visible. Most of the systems are competitive and they can’t afford to spend time and effort to improve (or even track) attributes which are not tracked by anyone. Usually, end customers involved in the transactions have visibility of only part of the entity which is getting optimized. Media and marketing ensure the same! So in the short run, the visible part gets optimized and everyone feels great!

However, in the long run, people start discovering side-effects of this short-term optimization on the limited part of the entity. Not only the invisible part gets left out of optimization, mostly it is the one that is sabotaged to pave the way for optimization on the visible front! Sooner or later after-effects of the unoptimized part add up to yield side effects and to create a non-holistic experience. Let’s see a few e.gs to understand this long-term loss.




A quick summary of our problem statement -

Reliance on just numbers leads to a nonholistic experience in the long run

A few e.g.s

An organic entity cannot entirely be defined by numbers	Yet we rely on numbers & limit our visibility of entity	Leads to non-holistic experience in the long run
A sales team member’s performance rating	Performance evaluation is limited to growing revenues and customer experience	Sales team focuses on high ticket size leads or easy customers only to optimize their evaluation. Teamwork, culture goes for a toss.(Please refer detailed case study at the end of this article)
The intelligence of a school child	Intelligence/future of a child is defined only by his scorecard	Kids compete for this objective rank in the class and take undue pressure on getting marks. They may not learn but cram leading to a loss in long run. A child gives up playing, creative projects, etc. which are not factored in his rank
Eggs that we eat	Eating any egg irrespective of poultry practices, provides the same pack of nutrients	We optimize the cost of the egg and the nutrition values keep getting sacrificed in the long run. Various polutry practics such as type of feed given to hen, condition/quality of hen, storage condition of egg, etc. affect the nutrition values. Many countries have cage-free, organic feed labels that subjectively define quality, however, marketing has ruined the actual subjective value
Pills for a headache	The faster a medicine acts, the better it is	Focus on high doses and other effective means to cut pain quickly. Side effects show up late such as low body’s defense to self-heal; not solving the real issue due to which body triggered a headache

In all the e.g.s we tried to encapsulate an organic identity into a few objective parameters and it led to unwanted long-term effects. While one may argue that this is due to limitations of numbers, but I feel that is just a reality we live in. If we invest in creating all the numbers, after a certain point it starts leading to impractical and complex web of numbers. We need to create our solution to this reality. We need to embrace the limitations of numbers and evolve our thought process to define entity better. This is where a subjective understanding of the entity can add value for long-term benefits.

Add subjective understanding to cover limitations of numbers; consciously!

While this may sound obvious to a few, I have seen many misses of using subjectivity in the day to day discussions and then facing long-term effects. Recently I recommended a movie to my wife solely based on its IMDB rating. I didn't even check other subjective attributes such as genre my wife likes, kind of social message movie gives, etc. Given this trend, producers will soon start creating movie scenes just to boost ratings and not caring about the overall aesthetics of the movie.

We need to consciously leverage subjectivity along with numbers. We should remember numbers don’t define an entity, but an entity is partially defined by numbers. So don’t optimize P&L to increase revenue but optimize organization such that it yields higher revenue and reflects in P&L. Define organization subjectively and not by just P&L. This subtle difference will make all the impact.

In essence, we should learn to use numbers and objective measurements as a tool to remove our biases and set a common ground and then build subjective understanding over it to create best solutions for a long run.

Let’s twist our earlier construct’s step -

Let’s go back to our earlier example. In case of performance rating of a team member, we used the weighted average of revenues and customer experience as a base tool to understand the team member’s performance. This removed all the prejudices and set a strong ground to benchmark the team member. Post this, we asked each manager to use his subjective understanding of other broad attributes to mark team members’ performance to the right bucket. Most of the folks who were rated for e.g. 3.6 (via weighted average score), landed up at bucket rating 4, but a few of them also landed at 3 and a very very few also landed at 2 (these were culture misfits who were blowing away discipline in the entire team). With this, we were able to create a better long-term output in terms of performance culture in the team.

[Please refer case study, in the end, to understand in detail.]

Applying subjectivity & dealing with its challenges

Use of subjectivity isn’t that simple. In the above example, we had discussions with individual managers to understand why would they mark a weighted average score 3.6 in the bucket rating of 2 or 3 or 4. These discussions were designed to be constructive and to be subjective. We had to train our managers to come out of the comfort of objectivity and gear up for such discussions. It took us some time but in the end, we had a team content with their ratings and aligned to perform holistically in the long run.

There are inherent advantages of pure number based logics and that is why people have an inclination towards them. To build subjectivity in our thought process, we need to overcome (a lot) of challenges. Let’s see a few inherent advantages of numbers over subjectivity and how can we solve challenges of subjectivity -

Inherent advantages of solutions based on numbers	Challenges of solutions with subjectivity & numbers
Neutral and doesn’t have biases	Subjectivity is the property of passion, beliefs, and values. Depends on biases of individuals and there can be no absolute truth-value assigned to it.
Automatically implied by logic	Requires logic to support it. Team members need to have trust in the last mile alignment
Skill needed: analysis	Skill needed: Leadership & communication over analysis
Easy to clean disagreements	Constructive discussions and leap of faith is required to clean disagreements
Group concludes easily based on logic	Group concludes with emotional cords backed by logic
Optimum for analysis over a large number of entities. For example, for a principal to judge the smartest kid, he would simply rank the scorecards	Optimum for specific analysis of an entity you are closely associated with. For example, a parent would build subjective knowledge of child over scorecard to help him improve further. Last mile decisions about an entity should be taken as per its specific needs only.

Subjectivity is worth it despite its challenges

My belief is that as one grows up on higher roles, limitations of number become bigger and you need to get a better handle on subjectivity to navigate further. That’s a challenging road but it is worth it as one grows in his roles and responsibilities. I will end the section with an excerpt from TN Hari -

• The general belief is that ‘Subjectivity’ is bad
• Any problem solving intentionally involves both subjective and objective components
• “extreme subjectivity”, “extreme objectivity”, “arbitrariness” is BAD
• A structured program & process with an appropriate balance of subjectivity & objectivity is GOOD

Key takeaways

1. Relying only on numbers will lead to non-holistic experience in the long run
2. Subjective understanding will add value in the long run
3. Subjective understanding comes with challenges but it is worth it!

Also to warm you up on my next blog, let me ask you what’s next?

Next blog teaser :)

Appendix - Case Study

Let’ take the case study of performance appraisals that we were designing for all the team members in TravelTriangle. I was wondering how to evaluate a team member’s performance rating.

An individual’s performance has many attributes. Let’s look at broad categories in sales team -

• Output based - monthly revenue targets, conversion % of leads and monthly % of happy customers
• Effort based - # of phone calls done every day and the soft skills while talking to the customers
• Other miscellaneous attributes - teamwork, initiatives outside the job, etc.

Let’s talk about various limitations involved in creating a perfect rating based on all these attributes

• If we include all the above attributes, it will be a long list of 20+ attributes. We might do a weighted average to come to a final rating to score team. However, it is very difficult for a team member to create a direction to focus his efforts on so many attributes. He will be lost and his net output will dip in all the KPIs. That’s the limitation of the human mind to absorb complex web of numbers for simple things
• Also to put the right score on a long list of attributes per individual needs high cost and time. There is certainly diminishing results of this cost
• It is impossible to put all the perspective in numbers. For example, earlier the ratings were a weighted average of revenue and customer experience. Each individual had optimized his processes to maximize these numbers. This had led to the overall bum of these metrics at the monthly level. So far so good. But on a deeper tinkering, we realized their efforts were misaligned for long-run benefits of the company. For example, in case of revenue optimization, pockets of sales team would 
• Put much higher efforts on high ticket size leads and will just not work on low ticket size leads. This led to a loss of the customer who would have come for a repeat with us in the future at high ticket size
• Put higher efforts on only those leads where customers’ demands were easy as they would naturally rate high on customer experience. This led to a loss of the customer who would require a little extra effort this time but would have been our advocates in future

Now I may re-engineer my KPI targets for revenue to include the above cases. For example, setting limits on the proportion of low ticket size: high ticket size bookings. This would solve this particular case. This kind of approach will have following cons - 

• It makes system complex, impractical and we start losing focus on high-quality metrics. A lot of times this leads to analysis paralysis even though it acts as an aphrodisiac to an objective mind
• Also, as we scale, such cases would keep coming up and in the end, our KPI re-engineering would always be a never-ending game. 
• A lot of these attributes are continuous (suitable for analog measurements) and not discrete (suitable for digital/excel reading based measurements). Hence it becomes tricky to accurately map values of such attributes in our discrete buckets. For example, to evaluate soft skills of a person, we had a weighted average of a few of its components (such as on-time response, patiently handling erratic customers, etc.) We decided to bucket team member’s performance on every attribute in just 5 practical buckets (viz: Super above target, above target, on target, below target and super below target. We can rate them from 5 to 1; 1 being the lowest bucket). However net weighted average scores would come in decimals and there would be another exercise to map them to the bucket. Let’s take some e.g.s if we simply round off 

Net Weighted score	Bucket mapped (by rounding off)
4.4	4 - Above target
3.8	4 - Above target
3.6	4 - Above target
3.4	3 - On target

The obvious issue here is that both team members who score 3.6 and 4.4 respectively, got mapped to the same bucket (4 - Above target). 3.6 and 4.4 were far apart by 0.8; which is a huge difference. So how do we solve this?

One approach could be to create more buckets to increase mapping accuracy - instead of 5, we can increase it to 10 or further to 100. However, this will lose practicality from the equation - people won’t have a clear understanding of how each bucket is different.

Due to all these limitations, it is near impossible to come to a perfect rating based on numerical exercise. Many organizations keep it simple and pick revenue and customer experience as broad metrics for the rating. Whatever rating comes from this decides the performance of team member entirely. We ignore all the perspective of performance not captured in the numbers above. What does it lead to?

• Most of the systems are very competitive and they can’t afford to do spend time and effort to improve (or even track) attributes which are not visible and no one is asking or tracking. As a result, even with good intent, other attributes gets neglected and harm the company and all the team members in the long run. Teamwork, outside job initiatives, and similar other attributes get neglected and ruin the culture in the long run.
• Simply rounding off ratings to the nearest performance bucket leads to discontent in the team esp the ones who get rounded off from the upper extreme. Many may forego it as a strike of hard luck but others will lower their performance a notch as they would still map to the same bucket

How we solved it

Let’s go back to our earlier example to understand the details. In case of performance rating of a team member, we should embrace the limitations of numbers and use a subjective understanding of the team member to complete the rating. For example, we used the weighted average of revenues and customer experience as a base tool to understand the team member’s performance. This removed all the prejudices and set a strong ground to benchmark the team member. Post this, instead of just relying on this to get to the final rating (rounding off it to the nearest rating bucket), we asked each manager to use his subjective understanding of other broad attributes to mark team members’ performance to the right bucket. Since managers knew their team members performance all year around, they were able to do this effectively. To remove biases of managers we asked them to set their base with objective scores first. Most of the folks who were rated for e.g. 3.6 (via weighted average score), landed up at bucket rating 4, but a few of them also landed at 3 and a very very few also landed at 2 (these were culture misfits who were blowing away discipline in the entire team). With this change, everyone would see performance as the complete entity - part of it via numbers and the other part via subjective understanding. Clearly, they were not relying on just the numbers to map performance but they were looking at performance holistically and numbers just as a tool that partially defines it.  We were able to create a better long-term output in terms of performance culture in the team.

Challenges of applying subjectivity

We had discussions with individual managers to understand why would they mark a weighted average score 3.6 in the bucket rating of 2 or 3 or 4. These discussions were designed to be constructive and to be subjective. We had to train our managers to come out of the comfort of objectivity and gear up for such discussions. It was a long road but in the end, we had a team content with the way they were rated.