Marginal gains are massively overlooked in legal service delivery. And it’s a shame. Marginal gains are about making small incremental improvements to a process (or product, ability or habit etc) which, when added together, make a significant improvement. But legal innovation often ignores it. Why? We’ll find out. We’ll also provide an easy intro to a few mental models that can identify and deliver these gains within your legal organisation, whether law firm or in-house!
What are marginal gains?
Consistently improving something by 1% isn’t particularly notable, let alone noticeable. But in the long run it can be meaningful.
- If you improve something by 1% each day for one year, you’ll end up 37 times better by the end of the year.
- Conversely, if you only get 1% worse each day for one year, you’ll decline down to almost zero by the end of the year.
If the principle sounds familiar, you may have heard of similar ideas, e.g. continuous improvement or Kaizen. Both continuous improvement and Kaizen have as their goals the ongoing improvement of products or services and the elimination of waste through data-informed iterative analytical improvements. Many of these improvements will be small, but when added together represent a large improvement.
It wins gold meals
The marginal gains theory was popularised by British Cycling‘s 2008 and 2012 Olympic cycling team. Prior to 2008, British Cycling was widely regarded as a “laughing stock”. In an effort to fix things British Cycling hired performance director David Brailsford.
Brailsford believed that 1% improvements across all aspects of cycling aggregated into a significant performance gain.
Under his guidance British Cycling implemented all sorts of tiny improvements:
Using antibacterial hand gel to cut down on infections, rubbing alcohol on bicycle tires for better grip, redesigning bike seats for greater comfort, extensive wind tunnel testing of bikes and racing suits etc.
Within 5 years after Brailsford took over, British Cycling won 57% of all road and track cycling gold medals at the 2008 Beijing Olympics, and at the 2012 London Olympics set 9 Olympic records and 7 world records.
During the ten-year span 2007 – 2017, British cyclists won 178 world championships and 66 Olympic or Paralympic gold medals and captured 5 Tour de France victories.
Quite a turnaround.
Big returns are the accumulation of small wins
Over time, the difference between allowing something to worsen by 1% each day vs. improve by 1% each day becomes extremely significant. Consider the below:
If you start the year with £100 of value and manage to improve that value by 1% each day for 365 days you 37x that value by year end. Over time, assuming the same 1% per day incremental improvement, that value balloons to a very large number.
Conversely, letting that value degrade by 1% each day quickly reduces that value to a very small number. £100 becomes £2.55 after one year of daily 1% declines! Thereafter it reduces further and further toward, but never reaching zero, although in real (as opposed to mathematical) terms, it hits zero just after 1 year.
As this graph demonstrates:
Which curve would you prefer to be on? The blue or the red?
Small is easier to start
If you break down a larger goal, problem or process into smaller steps, it is easier for an individual or organisation to measure, critique, brainstorm solutions for and implement winning improvements over time. As attractive as double digit improvements sound, they are often hard to contextualise and action, let alone start.
Competitive edge hides in the margins
As with the British Cycling example, focusing on 1% marginal gains can create a competitive edge. This is all the more so given how easily and often marginal gains are overlooked.
Car manufacturing, where similar principles of continuous improvement and Kaizen originate, paint a similar picture.
Toyota, pioneer of the Toyota Way and derivative principles (e.g. Six Sigma, Lean), used this attention to detail and marginal gains philosophy. In doing so Toyota quickly leapfrogged the US car manufacturing industry in the latter half of the last century, building a lasting reputation as a reliable manufacturer of superior and well priced cars vs. its overpriced and unreliable US counterparts.
Why marginal gains are wrongly overlooked
1% returns excite no one
Instinctively a 1% improvement to a product or process sounds too small a target. You’re probably aiming higher: 25% or 50%+ gains.
Put it this way: given the choice, would you prefer someone promising a 25% gain today or a 1% improvement each day?
You’d probably choose the 25% gain. But is that wise?
As we’ve shown above, ignoring marginal gains is at best, missing out on significant long term gains, and at worst inviting terminal decline.
Using an admittedly extreme example, if organisation A and B start with the same £100 of value, and A improves by 1% each day whereas B worsens by 1% each day over the course of 365 days, at the end of the year…
…A’s value is 1,481x B’s!
This is the same reason we don’t all have six pack abs: you have to work at them every day, via diet and exercise, to attain and then maintain them.
1% each day requires a long term mindset
The difference between being 1% better or worse each day is minimal at first. But as the above graph and example of organisations A and B demonstrated, these differences accumulate and widen significantly over time.
Most legal organisations are short-termist, and easily disregard the accumulated benefit of 1% gains over a year. But doing so is giving away a significant opportunity, or worse, inviting terminal decline if your processes and products trend 1% worse each day and a competitor’s improve 1% per day.
Legal (spin)innovation, FOMO and Shiny Objects are easier than 1% each day
Legal innovation suffers from three ailments:
1. Shiny Object Syndrome (SOS). Chasing the latest shiny #legaltech simply because it is new and in vogue, or overhyped in the press.
2. FOMO. Buying skills and / or software because of a peer firm’s flashy press release.
3. Spinnovation. Focusing on winning awards and issuing glossy press releases.
In each case, this is busy work. It has the appearance of innovation and is often quite easy, i.e. buy X, announce X… buy Y, announce Y and so on.
Now we have some logos for a pitch deck, our website or the next conference or media plug.
But did it change anything for the better?
Has the customer benefited? Is anyone in the organisation using these tools or applying these newly trained skills? Did they actually need them? Are they able to use them, and use them well? Does the organisation know what it already has and, if not, why not? Of the things it already has, how many are actually used by the organisation, or used poorly? Would there be a higher ROI if you made everyone expert users of Microsoft Office? (Check out this excellent boxset of how-to videos re Microsoft Office hacks by Crafty Counsel) Did you just buy something that overlaps 1:1 with another thing you bought last quarter that no one used?
Why is this?
Time spent asking these questions and listening to the answers is time well spent, and better spent vs. diverting resource to press releases, pitch materials and the never-ending consumption of new.
But what is simple to do is simpler to not do. And that is what usually happens.
All too often the legal market chases change via consumption.
But like buying a new outfit to make you feel good, albeit temporarily, would it be better to buy a gym membership and do the hard yards of diet and exercise to get in shape and look good in a scruffy t-shirt, or better still, naked?
But as with most things in life, there are few / no silver bullets, despite what vendors promise. Some of the biggest improvements are the accumulation of small wins made consistently.
As someone famous once said, a six-pack wasn’t built in a day (OK we made that up, but you get the picture).
So too with legal innovation, let alone any innovation! It can’t be bought, but only built. Brick by brick.
And this is where marginal gains slot in
The marginal gains technique approaches improvement like a marathon, not a sprint.
It’s not about running full throttle in a short burst, or buying the most expensive running trainers.
Instead, it’s having the patience to critically and continually assess your diet and form in order to improve, whilst maintaining stamina to keep going for the long term.
Focussing on marginal gains – and some techniques to help identify them – may uncover that you already have tools or expertise in your organisation that are under-utilised through lack of awareness, training, buy-in or otherwise.
Righting these issues can deliver gains without having to buy more things and spread fewer resources across yet more things.
Likewise, the very issues in your organisation that prevented your existing products or processes from being successfully utilised might also be the very same things that – left unchecked – doom to failure anything new, no matter how shiny or prized by your peers in their spinnovative press release!
How to identify, measure and improve via marginal gains?
In simple terms: map your process. Identify the pain points. Measure them. Analytically and empirically design, implement, measure and test solutions. Did it improve? Yes, great. If not, try again. Once you get it working, build in controls to ensure the improvements are maintained and new ones identifiable.
Where are small things eating away at your overall process efficiency?
If you want a more rigorous framework, you might wish to consider a time tested method such as the Six Sigma concept, DMAIC.
DMAIC stands for: define, measure, analyse, improve and control.
DMAIC is a five-phase data-driven strategy to improve processes. It is integral to Six Sigma, but in general, can be implemented as a standalone quality improvement procedure or as part of other process improvement initiatives such as lean.
We’ll try and explain the basics below (which is by no means intended to be an exhaustive or comprehensive guide to this subject).
Our aim is to whet your tastebuds and inspire your own research into these mental models and how you can apply them to your organisation!
The overall process
The fundamentals are relatively straightforward, so don’t worry, you don’t need to be (or necessarily hire) a Six Sigma Master Black Belt!
The overall DMAIC process looks like this:
Before we breakdown each stage, let’s make this real.
Marginal gains for NDAs
Below is an illustrative mapping of a typical NDA (non-disclosure agreement, aka confidentiality agreement) workflow involving a seller, a buyer and the buyer’s legal team.
In this scenario, the seller (light blue) needs to communicate some confidential information about the target they are willing to sell, and the buyer (dark blue) needs to receive that information in order to determine whether or not to bid for the target.
And of course, the legal team for the buyer will need to handle the NDA’s legal and compliance points.
The 3 horizontal boxes spanning left to right are swim-lanes, indicating which party completes the relevant action or task.
The process starts at the Start (top left) and ends at the End (bottom right).
01 Define 🧠
Define the problem to qualify the improvement opportunity, the project goals, and the customer (internal and external) requirements for a particular process.
Define the focus, scope, direction, and motivation for improvement. Are you trying to reduce turnaround times? Are you trying to reduce negotiation rounds? Are you trying to remove the need for NDAs altogether? Are you trying to make NDAs self-service for the business, i.e. so the Buyer can complete NDAs in part or in whole by themselves, and alleviate pressure on Legal?
Without defining these you can’t seriously search for marginal gains (or any gains) in solution to that aim.
Define the customer. Who are they? Complete a voice of the customer exercise to unpick their experience of the current state.
For instance, how does the Buyer experience the NDA workflow vs. the Buyer’s Legal team? Do those experiences align or diverge? If so, why and how so? How does the Seller find the process of collaborating on the NDA across the table? What satisfies/dissatisfies them about the current process? What would delight them?
Define the current process (e.g. as we’ve started to do above). What does the status quo actually look like? Intuitively stakeholders in a process will often react that “I know what I do, why map it?“.
This reaction is driven by intuition (i.e. I know what I do because I do it every day etc) but also a fear that they will be forced to justify what they do, and by extension their worth in the process. In reality, once stakeholders start mapping a process it quickly becomes clear that the number and complexity of steps exceed their intuition. Be alert to this objection, and pre-empt it by stating the objective is to optimise not criticise.
Understanding this alone can have a huge impact: it teases out the complexity but also gets people thinking about whether or not they need all of the steps and / or whether there might be alternatives. It’s a good way of seeding the idea that the way we’ve always done it might no longer be the best way to do it. It shifts people from thinking about their processes on autopilot, and more like an engineer trying to understand a complex system and reverse engineer a better one.
You can’t improve what you don’t understand.
The collaborative aspect also forces a conversation between all interested stakeholders, who might otherwise have unconstructive conversations, e.g. the Buyer moaning about the Legal team being the “Department of No” or “Dealbreakers not dealmakers” due to their slow NDA turnaround, or the Legal team moaning that the Buyers “never follow instructions” or “never provide the right information” etc. Getting them working together to solve a shared painpoint will result in a mutually beneficial solution.
02 Measure ⏱️
Measure process performance to quantify the problem.
Measure the current process. Taking the above process map, develop it into a value stream map by quantifying the time it takes to complete each step and the volume of each step (e.g. negotiation rounds) plus other metrics, e.g. the overall volume of NDAs per day, week, month and year etc. What does your baseline performance look like? Is it what you expected? If not, now you have a starting point to find out “why not?”.
Perform a capability analysis. How well or not does your process meet the necessary capability? Is the volume throughput of new NDAs (supply) exceeding the ability to turn them around (capacity)? If so why? How often are problems occurring in your process (and where)? What are their effects on the results?
Consider the Pareto Principle (also known as the 80:20 principle). In many systems, 80% of the positive or negative results will derive from 20% of the inputs and vice versa.
This can be a great focusing technique to hone in on the most problematic or positive parts of a process to either reduce or increase their influence respectively. Can you spot these splits in your current process?
If you are data-savvy, you might wish to analyse your data via a Pareto chart, which summarizes the frequency of problems or their causes vs. their criticality (and resulting impact).
03 Analyse 🔬
Analyze the process to identify root causes of poor performance (defects).
Perform a Root cause analysis (RCA) to uncover causes. It’s no more complicated in essence than: (1) determining what happened (i.e. your current process and what works / doesn’t work), (2) determining why that happened and (3) figuring out what to do to reduce the likelihood that bad will happen again. There will usually be at least 3 basic causes in play.
- Physical causes, e.g. the DMS being exceptionally slow to download / upload documents, introducing waste into the process (DMS Problem). Using old-school e-signature paradigms, such as printing, signing and scanning the NDA is slow and unnecessary (E-Signature Problem). Working that way also fossilizes the document as a flat image of a PDF that is totally unreadable by any contract review or reporting software (including shiny AI tools) without unnecessary and lossy optical character recognition to reconstitute the original text data (Data Problem).
- Human causes, e.g. the Buyer always forgets to tell Legal the basic context for the NDA such as the counterparty’s details, the Seller’s legal contact etc (Workflow Problem).
- Organizational causes, e.g. the Legal team’s playbook is too onerous and mostly unnecessary, resulting in significant time waste marking up NDAs or drafting new ones that the Buyer’s counterparties will reject and pushback throughout excessive negotiation rounds (Playbook Problem).
There are lots of sub-techniques to RCA. A particularly simple yet effective one is the 5 Whys. Ask “Why?” until you get to the root of the problem (usually 5 times is all you need).
For instance the below.
You can also branch off each reason returned by the answerer to delve into specific sub-issues and lines of enquiry, e.g. as above for the response “We can’t process them [NDAs] fast enough“.
After 5 Whys you can usually get to the root cause and begin to formulate a countermeasure that goes to the heart of the problem.
You might get to the root cause without 5 Whys, but you risk solving a higher-level issue that is a symptom of the root cause. Left unidentified and untreated you may solve that symptom, but another will likely pop up somewhere else in the process because the root cause remains intact.
04 Improve 🏗️
Improve process performance by addressing and eliminating the root causes.
Generate solutions by team brainstorming potential ways in which to solve the identified problems.
Critically evaluate these solutions. Do they solve the problems in the best possible manner? Consider applying the 5 Whys to the solutions to sense check that the solution does solve the problem (and why). What criteria matter most in solving the identified root causes? Should these be weighted? If so, how?
To focus this exercise you might use a Solution Selection Matrix.
To do so, brainstorm the selection criteria, weight those criteria and then evaluate them against their ability to solve the identified issues.
In the above Solution Selection Matrix (simplified), three solutions A, B and C are being compared and scored against x 4 criteria:
Easy to Implement, Quick, Problem Solution Fit and Customer Satisfaction.
Each of these is weighted as follows:
Easy to Implement (3), Quick (5), Problem Solution Fit (9) and Customer Satisfaction (9).
Everyone involved in the project votes, comparing each solution against each criterion.
E.g. Assuming a team of 10, each person possessing one vote per criterion, when comparing the solutions for ease of implementation:
- A received a score of 18, i.e. 6 votes x a weighting of 3
- B received a score of 3, i.e. 1vote x a weighting of 3
- C received a score of 9, i.e. 3 votes x a weighing of 3
Comparing this way you then sum across the criteria against each solution to determine the highest scoring solution, in this case A!
Then, using the above techniques (or something similar), select potential solutions for pilot. Ask yourselves:
- What are the smallest, most easily actionable ways to test these solutions quickly and effectively to generate feedback and more data about whether the change improves or degrades the process?
- How will you measure their impact?
As you test these solutions, measure the impact of these changes – do they improve or degrade the process vs. the baseline process you quantified in phases 2 and 3? If they improve, keep. If they degrade, discard and try again. If they failed to improve, consider re-applying phase 3 to those solutions to identify what remains wrong despite your improvement efforts.
Once you begin to improve your process and unlock your gains, marginal or otherwise, how do you sustain (control) these improvements?
After all, there’s no point doing this hard work only for things to revert back to the baseline performance or worse, degrade further.
05 Control 🚨
Control the improved process and future process performance.
In the Control phase the objective is to create a monitoring plan and continue to measure and revisit the success or failure of the process and identify emergent issues or areas for improvement.
Having done the earlier work, how does the team ensure the process maintains the gains and avoids new losses? Are there ways to build in controls that alert stakeholders to such issues and generally ensure performance maintenance?
A simple way to do this is to create an internal structure, policy based or, ideally automated, that tracks the measurements from phase 2 and 3 as part of an on-going basis.
If you can build in alerting mechanisms to flag performance increasing or decreasing, do so.
And don’t forget, reconnect with your stakeholders at an agreed cadence. Does everyone remain happy? If not, this is an indicator that you need to revisit the DMAIC process and continue your improvement.
The end result?
Hopefully, you’ll have a deeper understanding of your process, the people involved, the priorities and the quantum of any efficiency (or not) of your process.
In doing so, you’ll gain a clear understanding of where solutions might lie, and through analytical trial and error, improve the process toward your desired state.
Having done so, ensure to control the improvement in order to maintain performance, and to remain alert to future performance opportunities or performance degradations as and when they occur.
Investing this effort the first time around, it will be easier to repeat this process on an on-going basis.
Overall, you will unlock some marginal gains that add up to a significant overall gain, provided you continue to focus on improving processes on an on-going basis!
Remember 1% better each day is a 37 times improvement by year end!