We just explored way to properly
use your Demand Funnel to diagnose and fix problems with your Lead processes.
Using this diagnose-and-fix methodology, you should be able to get leads moving
and plug and any leaks in your process. This should create a positive lead flow
with negligible leakage or stuck leads and an “always-on” diagnostic to keep
you from developing more problems. So you’re good, right? Not so fast! You now
need to use the Demand Funnel infrastructure you’ve created to optimize your
process. Here’s how.
Understanding Optimization
Optimizations
makes some assumptions, so let’s make sure we have properly aligned on those.
1.
Optimization
assumes your process works. If your process is broken, optimization will not
help. It may make matters worse.
2.
Optimization
should be built in to your Lead Management process. It is not an optional thing
you do every once in a while for specific processes or parts of your process.
To be really successful, optimization needs to become a part of your Demand Gen
culture – something that is an expected part of the way everyone does their
jobs.
3.
Optimization
requires a framework and methodology. You cannot optimize against a moving
target.
We have already
covered many ways to perfect your process, so we’ll spend little time here
addressing that subject. At this point we need to assume your processes are
defined and properly implemented.
Building Optimization into your
process
If we think
of Demand Generation as a series of actions taken to achieve success against a
given strategy, we would normally execute these actions with a plan (as opposed
to random acts of marketing). Organizations manage those projects in a variety of
ways with a variety of tools – most often referred to as project management. The three pillars of project management are: 1)
Scope, 2) Resources and 3) Time. To build optimization into your project plan,
you are likely affecting all three of those pillars, with scope being the key. Let’s
look at a simplified example.
For every Demand
Generation project, no matter which tactics or media are used, major project categories
remain relatively constant. Starting with the idea that we would like to do X
by Y date, you may see a project plan similar to this:
1.
Initiation
and Planning: What do we want to do? Who is
our target audience? What defines success in Demand Funnel terms? How will we measure it?
2.
Discovery: What are the requirements,
resources and constraints to doing it?
3.
Ideation: Based on our discovery, how should we do it?
4.
Design: Based on ideation results, what do the
deliverables look like?
5.
Development: Based on
approved design concepts, build the deliverables.
6.
Testing: Based on the design
requirements, make sure the deliverables meet requirements.
7.
Deployment: Execute the
delivery of tactics to the target audience.
8.
Reporting: How did we do
against our plan?
As a Demand
Gen professional, you probably recognize project plans that look similar to
this. Within each of these major categories will be many sub-categories, each
dependent on the scope, time and resource availability. Of course, your
organization will have its unique twist to this structure.
What’s
missing? Optimization. Where should
optimization fit in this plan? Let’s take a look at a revised project plan
including optimization.
1.
Initiation
and Planning: What do we want to do? Who is
our target audience? What defines success in Demand Funnel terms? How will we measure it? How will we optimize it?
2.
Discovery: What are the requirements,
resources and constraints to doing it?
3.
Ideation: Based on our discovery, how should we do it?
What should be
optimized?
4.
Design: Based on ideation
results, what do the deliverables look like? What components of the deliverables should be
optimized and compared to other similar programs.
5.
Development: Based on
approved design concepts, build the deliverables. Build the optimization components.
6.
Testing: Based on the design
requirements, make sure the deliverables meet requirements. Make sure the optimization
components and tests work per specifications.
7.
Deployment: Execute the
delivery of tactics to the target audience. Implement the testing components.
8.
Optimization: Evaluate test components on the prescribed
cycle. Report and record results within the framework. Deploy winning
components.
9.
Reporting: How did we do
against our plan?
Notice we
only added one major step to the plan, but optimization has become completely
integrated into the planning, design and execution of the overall plan.
Optimization Framework
Optimization
on a particular program is great. Great
for that particular program. But how do we scale optimization so we can
leverage optimization learnings across Demand Gen programs for the entire
organization? For that, we need a framework.
A framework
is simply a method by which we can organize and execute our testing and optimization
such that results can be compared to other, similar tests. Think back to high school
science classes where you learned the Scientific Method. Your testing framework
should emulate this methodology to maximize your results and create the ability
to measure them across multiple programs of similar – and, in some cases,
dissimilar – types of programs. To get to an operational framework, let’s start
with objectives and work backwards.
Objective: Create 1,000 MQLs in Q1
Let’s break
this down. (Although your number is different, this is probably a very common
objective for Demand gen teams, so it will be easy to relate.) We have two real
options here. The first is to create MQLs from your unknown TAM – those contacts
outside your database, but fit your target market. The second is to convert
current Inquiries (those contacts already in your database who have previously
interacted with your brand) to MQLs. This sounds like the basis for an
optimization test, so what would we want to understand from this program test?
1.
Total Volume
of conversions: did more Inquiries or Unknowns convert?
2.
Speed of conversions:
Which converted faster – Inquiries or Unknowns?
3.
Cost of conversions:
Which converted at a lower cost - Inquiries or Unknowns?
As a result
of this optimization framework, we would begin to understand how best to apply
our marketing resources to the challenge of creating MQLs. However, we need to
dig deeper into optimization because we need to understand why these conversions
happen. Without that understanding, we could wind up applying the wrong tactics
for the wrong reasons. This is where our use of Scientific Method gains us an
advantage in optimization.
We start
with a hypothesis and devise a method to test that hypothesis against a
baseline. Let’s say our proposed program is designed to execute a very typical multi-touch
mix of tactics, including an email and a direct mail message. We have looked at
the previous arguments and decided to address current Inquiries in our database as our target audience, with the objective
of creating our 1,000 MQLs. It have been determined that conversions (via form
submit) from this program will generate sufficient scoring points to “auto-MQL”
any Contact already in Inquiry status. Let’s also assume our benchmark conversion
rate is 5%, so we require a total audience of 20,000 Inquiries for this
program. Our optimization objective is to convert at least 6%, beating our
benchmark rate by 20%. How, then should we design our optimization tests?
Let’s start
with our hypothesis. We believe offer messages work better than product
messages because our product is well-known and comparable to other products in
the industry. Price drives conversions.
The
conversion points for this program are:
Data and
list hygiene, along with good emailing practices, are the keys to deliverability.
We don’t test for that at a program level, so let’s assume static
deliverability for this program.
Open rates are a direct result of the
combination of sender and subject lines (for email) or packaging for direct
mail. We’re going to test two specific messages:
1.
Better Widgets, Better Price
2.
20% off Widgets until Friday.
Once the
email or direct mail is opened, we need to get the recipient to read far enough into our message to respond to our call to action (CTA). We’ve
already established that our products are well known, so we don’t need a lot of
description, other than why the reader should respond now. Here, we are going
to test two messages and two CTAs in a multivariate matrix to understand what
combination provides the best results. This test covers both the Read and Respond steps in our conversion path.
Message
1.
Our great
widgets do more for you.
2.
Our great
widgets are on sale.
CTA
A.
Better
Widgets, Better Price.
B.
20% Off this
week only
The matrix
is very simple, covers every possible combination, and looks like this:
1A 2A
1B 2B
Conversion happens once the recipient has responded, and is the
final measure of how committed the recipient is to your message and offer.
Whether the response is to purchase or to engage more deeply with your brand,
this commitment is the critical step. Do not assume that once your recipient
has clicked the CTA, he or she will go ahead and execute the conversion.
Abandonment here indicates you have not solidified your “closing message.” Let’s
try something to test that commitment here – I call it the bonus close.
For three of the four message/offer combinations, a sale price is
expected. Only one expects just a “better” price. Let’s align each of the four
combinations with a bonus close test.
·
1A expects
only a better price. Bonus Offer = 20% off
through Friday
·
1B expects
20% off this week only. Bonus Offer = 22% off if you order by Wednesday
·
2A expects an
unidentified sale price. Bonus Offer = 20% off through Friday
·
2B expects
20% off this week only. No Bonus Offer = 20% off through Friday
These tests will help us understand the best mix of offers and
bonus offers as they meet or exceed expectation when the recipient hits the
conversion form.
Change the conversation.
Optimization
is a conversation that needs to be a part of every program execution. As you
have seen here, it is a lot more than a few random tests thrown into the mix to
get a few extra conversions; it is an entirely new way of thinking about what
drives those conversions. The conversation should not be if we test, it should be how
we test. Smarter tests lead to smarter programs.
Notes:
Test against solid hypotheses to
provide scalable optimization.
Optimization is a culture, not an
action.
Test every conversion step in
your program process.
Are you all
in on optimization? I hope you are. One last thing to consider: what if one of your
tests goes wildly wrong? (This is not a bad thing – it is why we test!) Is
there a way to mitigate risk and not put your Demand Funnel in jeopardy?
Excellent question, and we’ll look at ways to do that next week in: Deodorant Goes Viral!
Test intelligently.
