Professional‎ > ‎Articles on-line‎ > ‎

Optimal Do/Tell Balance

posted Oct 20, 2011, 12:07 PM by Konstantyn Spasokukotskiy   [ updated Oct 20, 2011, 12:15 PM ]
The self-accountable product development teams and particularly entrepreneurial teams often have to balance where do they sink available talent and resources. If the balance isn't concisely defined and consensually maintained, there could be a significant drawback. For once it could result in overspending while reaching required milestones. Or it can instill a counterproductive egalitarianism. In any case the result can be a blame game or worse canceling the project. The positive side of explicit balancing is that one can actually plan required performance and talent demand in advance with great accuracy. It can reveal deficiencies and priorities for team development. hidden image

The classic microeconomics approach is to split the activity in out-capabilities and in-capacities, i.e. an ability of the team to wrap a product in a desirable package and an ability to actually produce the product up to the expectations. The outwards activity is in vogue referred as sales or Tell. The inwards activity is referred as Do. Herewith the required project resources are defined as:

Resources = Do + Tell

The created value, or the recognized by the market performance of the team, would be roughly proportional to the product of talent components:

Performance ~ Do*Tell

Do

The value of the product never ever progresses linear with applied effort. Most project managers know that a tiny fraction of the efforts defines success. Particularly, a typical well-managed technical project delivers more than 50% of value in the first third of the time. It is a logarithmic function. So the Do recalculated in output terms would transform last equation into:

Performance ~ log(Do)*Tell

If we assume that the performance evaluator is rational, knowledgeable and merit-striving then no further transformations of Do are required. An indifferent person is cynical. Such a person would have a product performance expectation level Expectation. She will depreciate the value of all all lesser results:

Performance ~ [exp(log(Do)/Expectation)/e]*Tell

There are special cases if product value becomes apparent strictly at the end of a project. Those are typically integration and MVP projects. Many truly innovative projects belong to this category too (see my presentation about Innovation Cycle, as well as note the performance metrics in the chapter "Effort Balance" below). Such projects can't be described by a logarithmic performance function. A step function or an exponential function would be more appropriate. The approach for a well-meaning thoughtful customer can be formalized as:

Performance ~ Do*exp(Do/100)/e*Tell

An approach for cynical customer would take following form:

Performance ~ exp(Do*exp(Do/100)/(e*Expectation))/e*Tell

Tell

Similar to the efforts for product creation the efforts for marketing the product don't translate into results one-to-one. There are two aspects here. 1) applied marketing tools, 2) diminishing efficiency of marketing efforts as the number of recipients grows. The later is similar to the case of logarithmic production output:

Performance ~ D*log(Tell)

The former represents three promotion methods:
  • Linear Marketing

    This is a message distribution method for saturation products. Doctor offices, competing grocery stores, nice to have, lifestyle products, etc. There is nothing exceedingly exciting about the product. The sales efforts don't differ from knocking on the doors and begging. The maximum Tell performance is conversion of 100% people you've pitched to, while facing one person after another. There are rarely well-meaning prospects in this category. The appropriate equation will look like:

    Performance = exp(log(Do)/Expectation)/e*Tell

    There is no adversary effect if the sales force reaches larger crowd. The costs of approaching prospects is high, so the effort will be wisely applied any way. The better your legs are, the better the outcome.

  • Power Marketing

    This method is applicable for mildly interesting products, when customers may mention the product once in a while during a communication with their peers. Or it can be applicable for saturation products if sales force uses active push tactics, like referral bonuses, coupons, etc. Assuming a well-minded prospect that multiplies sales effort by providing references Reference Multiplier, the appropriate equation is:

    Performance = log(Do)*Reference Multiplier*Tell

    The reference Reference Multiplier is a multiplication factor. It remains a low single digit number. Most likely it doesn't exceed 2. There is no significant adversary effect on the marketing efforts. Essentially it is still a job that leaves bones worn out.

  • Viral Marketing

    This method applies to sensational products and skillful marketing teams. Rumors and other guerrilla marketing instruments may create a wave of "uncontrolled" referrals. The description of marketing performance for this method is roughly similar to the logic in Metcalf's Law:

    outcome ~ nodes*(nodes-1)/2

    The outcome asymptotically equals nodes2. On the other side of the coin, viral marketing uses broadcasting instruments that demonstrate diminishing return if the size of target group increases. The appropriate total performance can be -best case- described as follows:

    Performance = log(Do)*log(Tell)2

    wost case:

    Performance = exp(log(Do)/Expectation)/e*log(Tell)2

Effort Balance

Almost every project team has limited total amount of efforts as described above by the resource equation. Assuming the total amount equals 100%:

100 = Do + Tell, 
Tell = 100 - Do

The question is how to distribute the efforts among the Do and Tell activities. There are two tasks while balancing efforts: 1) efficiency, 2) effectiveness.

The efficiency task seeks the maximum productivity, gain that is obtainable by a given work capacity. Assuming there is no penalty for bending discrete nature of labor units (product producer doing marketing job and vice verse), we can obtain a performance curve for different Do & Tell models. The curve will give us optimal efficiency balance at the point of maximum performance (see chart, vertical axis). Do value can be read from the chat below, horizontal axis. Tell value can be calculated as stated above.

The actual pike value suggests maximum relative performance for different approaches if calculated as a Performance1/Performance2 ratio. By comparing to the maximum yield (here small group viral marketing, i.e. Performance2) we can estimate effectiveness of other approaches.

efficiency balance at highest point
  • Small Group Viral Marketing
    Do 15%, Tell 85%, Performance 19750 = 1 unit

    A micro-scale viral marketing is the most effective marketing method. It is about 40 times better than the next closest method. If you can precisely identify your target group, it's the cheapest way to get desired results. One needs to dedicate approximately 15% of resources to manufacture a product and 85% to market it.

    Interestingly, in the USA only 18% of GDP is produced by manufacturing sector (all sectors that remotely produce goods 25%). It seems, the country is well-segmented in "little villages". The available communication instruments are effective to say at least.

    For integration projects: Do 41%,Tell 59%, Performance  79122 = 4 units

  • Linear Marketing
    Do 23%, Tell 77%, Performance 73 = 0.004 units

    This method is the most attainable. It can be practiced by non-professionals. Many purely technical startups, particularly B2B and nice to have B2C, practice it. The efficiency is quite fair, considering that material expenditures can be quite low - an attractive target for the Lean Startup movement.

    Integration projects: Do 0%, Tell 100%, Performance 37 = 0.002 units
    With optimal Do at zero here is a proof that cynicism kills normal business.

  • Power Marketing
    Do 24%, Tell 76%, Performance 483 = 0.025 units

    Power marketing can deliver second best effectiveness. It is 6 times better than the low-key linear marketing. For the best efficiency the manufacturing efforts should consume about a quarter of all resources.

    Integration projects: Do 62%, Tell 38%, Performance 3222 = 0.16 units

  • Viral Marketing - thoughtful attitude
    Do 36%, Tell 64%, Performance 62 = 0.003 units

    Integration projects: Do 74%, Tell 26%, Performance 606 = 0.03 units

  • Viral Marketing - cynical attitude
    Do 38%, Tell 62%, Performance 19 = 0.001 units

    A full-scale viral marketing -no matter what is the perception mode- isn't particularly effective. This is also a signal that nothing goes by itself. Great buzz requires a lot of guidance and nurturing. It costs a lot. An advantage is distant (non-face-to-face) leads processing. It can be used by robots as practiced in web businesses. Another advantage is a flat top. Viral marketing approach is tolerant to management mistakes, while doing resource allocation and project progress management. This is often an enabler for teams that don't have sufficient management expertise.

    Project teams want to add significantly more bells and whistles into their open beta products ~ 2.5 times more than closed alpha, i.e. comparing to small group viral marketing; and 1.5 times more in comparison to pet projects, i.e. linear marketing. The next chapter "Effort Costs" will give you an idea what it means to have a more advanced product from get-go.

    Integration projects: Do 68%, Tell 32%, Performance 9 = 0.0005 units
The performance distribution curves for in-core innovative or integration projects are on the following chart.
innovation performance in relation to production efforts
An important lesson is that innovation projects need 2-3 times more prefabricated substance before roiling out with any marketing approach. Given that innovative products generally require more intense marketing than incremental (anticipated) improvements, the lower marketing share for innovation products is actually larger in absolute terms. Assuming marketing costs are equal for incremental and breakthrough improvements, the optimal investment in innovative technology sans marketing should be at least:

Investment Multiple = (Dobreakthrough*Tellincremenal)/(Doincremental*Tellbreakthrough)

That multiple is 5.2 for power marketing, 3.9 for viral small groups, 5.1 for viral marketing to thoughtful people, 3.5 for viral marketing to general population. So much about the notorious micro-VC argument. Sufficient seed financing, capital formation at very early stages is probably the most important thing to keep entrepreneurial America afloat.

Effort Costs

Besides effort balancing the next most interesting questions in this ball park are: 1) what are the minimum required skills to accomplish a project, and 2) what are the tolerances for resource allocation that wouldn't jeopardize the project, i.e. acceptable range for management mistakes. With a little help, using a couple of industry specific metrics, these questions can be reliably answered.

First essential variable is Productivity Increase Multiplier that is capable to attract free floating capital. There was an estimation for the U.S. market in my presentation about Financial Bubbles that had said 5 times (>500%). Similar method can be applied to any other country. The data will be based on the particular national stock and business practices at the time of analysis.

The productivity increase can be supported by both superior soft skills of the business team (companies with service profile) and superior technology that provides functional gains or/and costs advantages to consumers (asset profile).

In case of service there is no much variance. A sales person can vary her work intensity from 0% to 100% and modulate it by selecting an appropriate marketing approach. Despite significant gains/losses that occur by preferring a particular marketing approach, the modulation isn't freely available. It is limited by the product's nature and availability of complimentary resources that the talent has to leverage -all largely exogenous factors. For example, if there is no budget for referral bonuses and promotional material the power marketing approach would necessarily default to linear marketing. Sales approach can impact customer engagement. For the only case when it matters - viral marketing - the difference in parameters is negligible and the real-life difference in performance is less than 3 times. It is safe to assume that a Tell-talent is primarily characterized by her ability to utilize given time productively, while avoiding mission impossible endeavors. The amount of productively spent time can be measured as a percentage point of the entire time resource. Summarizing: the sales people talent has a narrow effective bandwith from 0% to 300% of standard sales performance.

Similarly, the sophistication of created goods and assets can be liked to time utilization of an average engineer. Bright minds aren't necessary if standard combinatorial product development can do the trick. But in case the Do*Tell team-performance doesn't live up to required level, an inspired engineer can invent. Unlike sales, a technical talent would be able to add a magnitude at the Do-scale. Though, this immediately will throw the project into the in-core innovation basket, hence shifting all project parameters and making it harder to sell.

Assuming that the standard settled-down industrial complex has final productivity proportional to 100% of human time utilization, i.e.

Productivity ~ N*100

where N is the number of employees. The project team's productivity has to be:

Productivityproject ~ Productivity Increase Multiplier*N*100

If we have two primary competences Do and Tell, while each can be up to 100 productivity units or per cent, then the outstanding project's performance has to be at least:

Performance = Productivity Increase Multiplier*(Do + Tell)

The minimum required skills level can be calculated by solving the equation:

Productivity Increase Multiplier*(Do + Tell) = f(Do)*g(Tell)

where f(x) and g(y) are appropriate effort-to-result functions as described in the chapters "Do" and "Tell".

This equation has no practical solution for linear marketing. The maximum performance level at 73 units is by factor 14 less than required. Significant ingenuity is essential to make such projects attention worthy.

Assuming that the project team has all favorable conditions and acquired top-notch sales talent (100% performance and the performance boost at maximum = factor 3), the project performance will not exceed 700 units. In plain English: an extraordinary sales guy alone will not cut it on incremental improvements. The engineering team has to be 1.4 times more productive than a good engineer. Should the engineering team be 7 times more efficient, the marketing team has still to exceed expectations by at least 2.5 times. The entire productivity function that considers minimal requirements for linear marketing approach looks like following:
linear marketing, minimal requirements and productivity
Values of Tell over 300 = 3x average productivity at 100% efficiency, i.e. no-mistakes, dedicated employee, as well as Do over 1000 = 10x average productivity, or technology efficiency at the moment of market entry, are not realistic. There is just a narrow strip where an incremental improvement project can be successful. The success means reaching EVA at the level that is attractive for all stakeholders in free-market economy.

The skills function for viral marketing of incremental projects has no solution whatsoever. No effort will ever create sensible EVA. If you know such projects, they are a wealth redistribution or an ideologically painted highway robbery. It seems, any incremental or lowest-costs innovation project isn't a practical solution for positive wealth creation. If not counting for small scale business, as in case of small group viral marketing, it is either impractical or outright impossible. 

The skills function solution for viral marketing of in-core innovative projects is as follows:
The skills function solution for power marketing of core-innovative projects reveals that it is quite effective method to process projects. Despite high upfront capital requirements, the risk is low. The projects can be processed by ordinary humans. A realistic long-term 50% work intensity for both technical and marketing team members would be sufficient. It explains why rigid corporations, having minimal incentives to innovate, are actually successful and have relatively low project fail rate. The ability to subsidize market entry, i.e. to provide significant upfront incentives for committed distribution partners, pays off.
power marketing, truly innovative projects, performance requirements
To obtain real talent productivity metrics the reads form the charts above have to be related to Typical Employee Productivity metrics across the industries - the second industry specific variable. For example, it is common for automotive suppliers to have a revenue per employee in the range of $200k annually. In the IT sector this metric is in the range $300k to $1.3M. Most likely an average technical IT-guy would qualify for the genius role on a project in automotive industry. At the same time, an average IT-sales guy would be helpless in automotive industries, where there is not enough profit to sponsor large capital intensive break-through and promotional programs as done in web. The necessarily low-key marketing approach will demand much more refined individual skills to lead the projects into fruition.    

Summary

The theoretical contemplation shows that the decision making in resource allocation for innovation projects is complicated. Depending on how one restricts the thinking plane and sets of important variables, there are sub-optimums in resource allocation. I've seen a lot of failures when professional managers clearly fall for those sub-optimums AND were defending their straight as well as just optimal decisions.

Nonetheless, theory is just that a tool - not a master. A master beats theories by being even more sophisticated. Please write back if you have positive cases when this theory didn't work or was too exclusive while considering important phenomena in project management.