Today something complex, advanced, that is most applicable to those who are at the edges of spending money, and thus have an intricate web of internal and external teams to deliver customer engagement and business success.
The Marketing Industrial Empire is made up of number of components.
If you consider the largest pieces, there is the internal (you, the company) and the external (agencies, consultants).
If you consider entities, you’ve got your media agency, your creative agency, your various advertising agencies, your website and retail store teams, your analysts, marketers, advertising experts, the UX teams, campaign analysts, fulfillment folks, the data analysts who are scattered throughout the aforementioned entities, the CMO, CFO, and hopefully your CEO. And I'm only talking about the small portion of your existence that is your marketing and analytics.
Whether you consider the large, simplistic perspective (internal – external) or the more complex entity view, it’s really easy to see how things can become siloed very quickly.
It’s so easy for each little piece (you!) to solve for your little piece and optimize for a local maxima. You win (bonus/promotion/award). It is rare that your company wins in these siloed existence.
That’s simply because silos don’t promote consideration of all the variables at play for the business. They don’t result in taking the entire business strategy or the complete customer journey. Mining a cubic zirconia is celebrated as if it is a diamond.
Heartbreakingly, this is very common at large and extra-large sized companies. (This happens a lot less at small companies because of how easily death comes with a local maxima focus.)
So how can you avoid this? How do you encourage broader, more out-of-the-box thinking?
This might seem simplistic, but sometimes it helps to give things names. Naming things clarifies, frames, and when done well it exposes the gaps in our thinking.
Today, I want to name two of the most common silos in large and extra-large companies, in the hope that it’ll force you to see them and subsequently abandon siloed thinking and solve for a global maxima.
2. The Business Ecosystem: Active Engagement.
Name abstract ideas, draw pictures, deepen appreciation, take action.
Could not be simpler, right? :)
Let’s go!
The Advertising Ecosystem: Passive Consumption.
I'm randomly going to use Geico as an illustrative example because the frequency at which they are buying ads means that every human, animal, and potted plant in the United States has seen a Geico commercial at least once in the last 6 hours (contributing to Geico’s business success).
Typically the ads we see are the result of the external creative and media agencies, and their partners in the internal company team/s.
Geico purchases every kind of ad: TV spots, radio ads, billboards (OOH), digital displays (video, online,– social media), print (magazines, newspaper, your cousin's Christmas letter), and so much more.
The teams naturally gravitate towards optimization and measurement that spans their individual mini-universes.
Was that a great ad? Can we test different spending levels in that market? What is the best way to get people to remember the delightful gecko? Can we automate the placement of display ads based on desired psychographics?
Did we get the TRPs that we were shooting for? What was the change in awareness and consideration? What was the reach/frequency for the Washington Post? How many impressions did our Twitter ads get, and how many people were exposed to our billboards?
These are important questions facets of, and delivery optimization of, the advertising. Questions like these, and adjacent others, tend to drive the entire lives of creative and media agencies/teams. For entirely understandable reasons. Siloed incentives delivering siloed local maxima results.
I cannot stress enough that these results can be positive (for the ad business and, in this case, the sales of insurance products). And yet, as a global maxima person it does not take a whole lot of effort to see a whole lot of opportunity if both the siloed incentives can siloed execution implied by the above questions can be changed.
Here’s an incredible simple way that every human seeking global maxima can look beyond the silo: “So, what happens after?”
As in, what happens after the finite confines that are the scope of my responsibility/view?
To see that, the first step is to paint a picture that illustrates the current purpose (your silo), and then give it a name.
Here’s that picture for the example we are using, and the name I gave it is “passive consumption.”
Over 90% of advertising is passive consumption. This means that the ad is in front of the human and they may see it or not see it.
Even on the platforms where interactivity is at its very core (Instagram, Facebook, YouTube, etc.), almost all of the advertising does not elicit any sort of interactivity. If you look at the percentages, almost no one clicks on banner ads, a small percentage on search ads, and you need only speak with a few people around you to see how many people actively engage with TV ads vs. run to the bathroom or pull out their mobile phone the moment forced-watch TV ads come on.
Keep in mind, this is not a ding against passive consumption or the hard work done by Geico's agency and internal teams. Blasting ads on TV does cause a teeny tiny micro percentage to buy insurance – a fact provable via Matched Market Tests, Media Mix Models. The teeny tiny micro infinitesimally small number of views of brand display ads will cause outcomes. (Hold this thought, we’ll come back to that in a moment.)
So, what is the passive consumption challenge?
First, how far the vision of the creative and media agencies/teams will see (thus limiting success – global maxima). Second, trapped in the silo the vision for what will be measured and deemed as success.
The first is heartbreaking. The second ensures the death of any long-term impact.
Let me explain.
With over 90% passive consumption…. Well, passive… Smart media and advertising agencies/teams will primarily use post-exposure surveys to measure awareness (what companies provide car insurance) and consideration (which brands you would consider).
The brilliant agencies will also measure elements such as purchase intent (how likely it is that you'll consider Geico as your next car insurance provider) and likelihood to recommend (how likely is it that you'll recommend Geico to your family and friends).
All of these metrics will cause surveys to be sent via various mediums to people who've seen the TV ads, the banners on Facebook, and the video ads on YouTube. And a subset of users who were not exposed to the ads. Usually, there is anywhere between a few hundred to a thousand survey responses that will end up providing a statistically significant sample.
The scores from these responses are presented in weekly, monthly, or quarterly meetings. Segmented by marketing activity, they are the end-all be-all justification for media spending. Snapchat increased aided awareness by +23%, let us spend more there. Or, billboards in Georgetown and Austin shifted purchase intent by +2%, we should triple our spend in Chicago.
Every measurement and optimization initiative is based on this cocktail of metrics. Thus delivering a positive, but local, maxima.
Even the next best innovation in media will be based on results from the same metrics cocktail. Thus delivering a little more positive, but still local, maxima.
Why not global maxima?
Because success is determined by, innovation is driven by, measurement that is self-reported feelings.
That name captures the actual thing that is being measured (feelings) by the metrics above, and where the data comes from (self-reported) after being exposed to our advertising.
This will help your company, your agencies, understand limits. Limits in terms of what’s happening (mostly, passive consumption) and what data we are looking at (all post-exposure and self-reported).
Limits in measurement that incentivize solving for a local maxima.
Let me repeat one more time. Passive consumption measured by self-reported feelings does drive some success – else Geico would not be the financial success it is. In the short-term some campaigns are trying to drive long-term brand influence or causing a shift in public opinion or simply to remind people your brand still exists as a choice. All good. Self-reported feelings are wonderful. Appreciate that even in those cases where you are not trying to drive short-term sales, if all you have are feelings converted into metrics… You are limiting imagination.
An obsession with just passive consumption by your agencies and internal teams delivers 18 points of success. I’m saying if you think global maxima, remove limits, you can do 88 points!
The Business Ecosystem: Active Engagement.
Getting those additional 70 points success requires breaking the self-imposed creative/media/advertising silo and caring about the human behavior if people lean-in instead of passive consumption – when they take an action (a click, a phone call, a store visit).
Time to draw another picture, and give this behavior a name.
I call it… drum roll please… Active Engagement!
Some people, between 0.01% to 10% (so rare!), who see Geico’s online ads will visit a Geico retail store or Geico's website.
People are actually doing something. They are walking into your store, talking to an agent, picking up the literature, calling you on the phone, clicking on to your site, watching videos, comparison shopping, and more. This is all human behavior that your tools can report for you.
A small percentage will end up buying insurance – mazel tov! –, providing perhaps the most valuable data.
The lucky thing about active engagement is that, in addition to self-reported feelings, you also get tons of highly-useful quantitative data representing human behavior.
I call this type of data: Observed Human Behavior.
If you are a part of an creative, media, or an internal company team, you have two powerful issues you can solve for: passive consumption (happens most of the time) AND active engagement (happens some of the time).
Likewise, you can seek to understand performance using self-reported data where the people reflect on how they feel, along with behavior data that represents what they actually do.
The combination of these two factors deliver the much needed Global Maxima perspective.
That is how you shatter silos. The creative agency has to care about how ads perform in their labs, in the real world, and what kind of online and offline behavior the creative is driving (end-to-end baby!). The media agency has to care about the creative and where it needs to get delivered (recency, frequency FTW!), and the bounce rate (70% ouch, 30% hurray!) and profit from each campaign. The retail experience team, the call center delight team, and the site experience team will break their silo and reach back into understanding the self-reported feelings data from the media agencies and the ideas that lead to the creative that delivered a human to them.
Everyone cares about the before and after, solving for the overall business rather than their little silo. Passive consumption plus active engagement equals global maxima. Or, self-reported feelings plus observed human behavior equals global maxima.
: )
Here’s a massively underappreciated benefit: It also encourages every employee – internal and external – to take full credit for their impact on the short and long-term effects of their effort.
It is rare to see this happen in real life, even at top American and European companies.
What’s usual is to see the three silos between creative agencies, media agencies, and company internal team. There is usually further sub-segmentation into passive consumption teams (also lovingly referred as brand agencies/advertisers) and active engagement teams (performance agencies/advertisers). The further sub-sub-segmentation into products and services (depending on the company).
They then quickly fall into their respective measurement silos, solving for the local maxima.
Change starts with naming things and drawing pictures. Gather the key leaders at your company and agency partners. Show them passive consumption and self-reported feelings along with active engagement and observed human behavior. Talk through the implications of each picture. Ask this influential audience: What can you contribute to when it comes to breaking silos?
I have yet to meet a single company where simply drawing the picture did not result in a dramatic rethinking of focus areas, responsibilities, and ultimately priorities.
Accelerating Success: Five Quick Changes.
Once you have that discussion, what should you do to truly cause a significant change in behavior?
Five Es form the core of the strategies that I end up using (please share your's via comments below). They are:
1. Expand the scope of data your employees use.
For the people who buy your television ads, include both store and website traffic data. Break the shackles of GRPs and Frequency.
For people buying your display ads on Facebook, include page depth, bounce rate, as well as micro-conversion rates for those campaigns. Break the shackles Awareness and Views.
For people buying your videos ads on Hulu, complement Hulu's self-reported feelings metrics with user behavior and conversion rates.
And continue going in this fashion.
2. Expand the incentives structures for your employees.
Most marketing employees, both internal and external, undertaking passive consumption initiatives are rewarded for cost per TRP, effective reach, awareness and consideration increases, etc. Whatever this bucket as an employee incentive, it can stay.
Consider adding one or two KPIs from active engagement. For example: Store visits, phone calls (as a result of that increase in consideration). Website visits, loyalty, micro-outcomes, and 25 other easily-available observed human behavior metrics are available to you pretty much in real-time.
For people who own responsibility for your stores, call center and website, take a metric or two from passive consumption and make it a small part of their incentive structure.
People respond to what they are compensated with, or promoted for. Use it to solve for a global maxima in the company and its customers.
3. Expand the time horizon for success.
This is really hard.
You buy 100 TRPs, it’s expensive, and the executives tend to start badgering you for immediate results.
The problem is that self-reported feelings data takes time, and since at least 90% of passive consumption leads to no immediate active engagement, all this does is incentivize bad behavior by your agencies and employees. Long-term objectives are thrown onto the chopping block and long-term strategies are judged on short-term success – which immediately ruins the campaign’s measurement. Oh and the audience being bombarded by your ads that are trying to deliver short-term outcomes from long-term creative and campaigns… They despise you because you are sucking, they can see that, and they instantly realize your are wasting their time.
No matter how much your wish, a Chicken won’t birth a Lion’s cub.
If you want short-term success, define the clearly as a goal, pick the right short-term self-reported feelings metric and observed behavior metric, now unleash your creative agency and their ideas (on that short-term horizon), then plead with your media agency to buy optimal placements, and ensure the retail/phone/web experience is not some soft and fuzzy experience, rather it is tied to that clear goal and success metrics. Sit back. Win.
If you want long-term success… Same as above, replace short with long. How amazing is that?
4. Expand the datasets that teach your smart algorithms.
If you’ve only visited this blog once in the last 12 months, or read just one edition of my truly amazing newsletter :), Marketing <> Analytics Intersect, it is quite likely I have infected you with the passion to start investing in machine learning in order to bring smart automation to your marketing and user-experience initiatives.
If you are following my advice, make absolutely sure that you are not training your algorithms based solely on passive consumption, self-reported feelings data. It is necessary, but not sufficient.
Rich observed behavior data will provide your algorithm the same broad view of success as we are trying to provide the humans in #2 above. In fact, the algorithms can ingest way more data and complexity. Thus allowing them to solve for a super-global maxima compared to our humble abilities.
Every algorithm is only as smart as the data you use to educate it. Don't short-change the algorithm.
5. Expand leadership comfort level with ambiguity.
For your TV efforts, there are limits to what you can measure. You have self-reported feelings data, and usually that’s about it. If you have a sophisticated world-class measurement team, you may be running some controlled experiments to measure one or two elements of active engagement observed human behavior data.
For YouTube or Hulu on the other hand, you’ll have additional self-reported feelings data, and if you follow my advice today, plenty of directly-causal observed human behavior data at your disposal.
Get very comfortable with this reality, and execute accordingly.
When some executives are not comfortable with this reality, they typically end up gravitating towards the lowest common denominator. Even in regards to strategies where more is possible (digital), they just end up using self-reported feelings data for everything.
I do understand why this is; executives are pressed for time, so the executive dashboard needs only one metric they can compare across initiatives. This instantly dumbs-down the intelligence that could help contribute to smarter decisions.
Kindly explain this to your executives, share with them the value of being comfortable with a little ambiguity that comes from using the best metric for each initiative type.
We can achieve smarter global maxima decisions if we just use different metrics in some instances.
The larger the company, the harder it is to solve for a global maxima. Companies need command and control. Companies worry that people are going to run wild in 15 different directions. Companies need to reward an individual, that means creating a finite role that can be defined and measured at a small level. Companies add layers upon layers to manage. Companies create org clusters (divisions). And, more.
Every one of these actions forces a local maxima. Every human can see their few pixels and have no idea what the image looks like.
Even if then the company progresses little by little, they’ll run out of luck one day. Worse some nimble small company – that does not yet have to worry about all of the above – will come eat your breakfast first, then dinner and then lunch.
The lesson in this post applies across the entire business, even if in this instance it is applied to marketing and advertising.
Paint a picture of what the local maxima execution looks like in your division – or better still company. Give these pieces a name. Then, figure out, like I’ve done above, what the connective tissue is that’ll incentivize global maxima thinking and execution.
Carpe diem!
As always, it is your turn now.
In your specific role, are you solving for the global maxima or a local maxima? How about your creative and media agencies? Your internal marketing or product teams? Has your company done something special to ensure that teams are considering both self-reported feelings and observed human behavior? Is there a magic metric you feel that’ll encourage each piece of the business success puzzle to solve for a global maxima?
Please share your wisdom, tips and secrets to success via comments below.
Thank you.
Via 
Love the 5 Expand recommendations.
In my experience all of them apply when it comes to solving a broad spectrum of challenges when we try to expand data's influence. I would probably rate #2 as #1. I've observed it as the root cause for most problems that I bump into in my work. With consideration of what's coming quickly, #4 would likely be my #2.
Great post, thanks Avinash.
Hallelujah! You've perfectly captured the worldview that most media agencies bring to the table in my experience. A large part of this is what the client is asking for, that is what sets scope. This is where your emphasis on a company's internal scope plays an important role. They ask for GRPs then they'll get GRPs back on a pretty slide. If they ask for three-second video plays without thinking about it, that will drive the campaign's execution and the success metrics.
That said most agencies don't have internal mechanisms to train people for the type of thinking you've outlined. Legacy, lack of access to data, skills being hired into media roles, agency retainer or payment structure, all play a role. None of this is desirable. It might not ensure that the agencies survive the upheaval that we all know is coming. I plan to use your two pictures in our agency to start a dialog.
I find this to be true as well. The scope of the work initially signed by the executives usually shape the data collection and analysis long before the media team/marketing team even get their hands on it. I also find that clients often don't have the mechanisms in-house to know what data to look for and optimize for. Thus the agency does what the client asks for. In my opinion, there's needs to be an educational component between agencies and between agency-client so that everyone on board can be headed for the global maxima.
I also find that payment structures often get in the way as well. Ideally, when the initial project is being outlined, all of this will begin to be taken into consideration.
Jules: I could not agree more, thank you for drawing out this nuance.
I expect the Agencies to have a broader view – they work across a wide spectrum of clients -, for them to lean-in when the initial scope of work is being set. If they help bust silos, they will only add more value to the client, thus ensuring more agency business.
That said, clients need to be smarter ultimately and that includes smarter payment structures. Absolutely!
Avinash.
This is such a refreshing change from what the Analytics teams and thought leaders discuss. A lot of talk devoted to metrics, dashboards and tools. Close to none devoted to these higher altitude yet hidden issues.
Thank you for consistently bringing up issues that matter the most. Love the drawings.
Do you have any tips for someone young in how to influence senior leaders on Global Maxima? I am quite low on the totem pole.
Andreas: I recognize how hard it is for the tail to wag the dog. :)
Here's a blog post with elements that create a deeper influence for data: Five Key Elements For A Big Analytics Driven Business Impact.
Please reflect on #2, #3, #4. If you have an opportunity even consider #5.
The other tip: Walk, don't talk. Most Senior Executives respond to things they can see rather than words they hear. So, mock up a new dashboard with real data. In your analysis, look at metrics they need vis what they want. Add savvy elements like Goals, Benchmarks, Profit. Each will earn you a small amount of credibility, and soon you'll have a mountain!
Avinash.
My product brain makes me think that all those large companies can afford to have a huge common dashboard that can help report everyone's inputs, action and output metrics laid out against an X axis of time on a common interface where everyone can see and try to understand what things are working in silos, how best they can be synced in order and how the numbers are showing up.
In short everyone sees everyone and work together as one. Easier said than done.
Amit: You are right. Easier said than done.
I find that when companies are that big, it is nearly impossible to have one dashboard capture the relevancy of all the pieces. Due to size, things get aggregated up. Big chunks of people and orgs get fuzzy very quickly.
One solution for this is having a clear line of sight. Relevancy then increases as you go down. I cover this a bit in this blog post, section on Cascade the analysis and responsibility for data.
The other solution is to have fractional ownership directly of the macro KPIs. This works as well.
Avinash.
This one could not have come at a better time for us. I am struggling to convince our IT department to move away from the provably wrong strategies that they are undertaking for our company – and always have – yet they are not budging. After reading this post on Monday I dug deeper and as you mention, in their silo what they are measured against only applies to a narrow view. If we have to have any chance of success, this has to change.
You've given me some ideas on how to go about doing that. You rock.
I have a question about one of your tips: Expand the scope of data your employees use.
Is it not more important to focus on 1-2 KPI's (outcomes) than expand the scope of data to break shackles of vanity metrics?
Maybe they will not use any KPI to define succes of their Facebook/Hulu campaign?
Gerard: Great question.
If you pick good KPIs (good list here: Analytics Tip #26: Every Critical Metric Should Have A BFF!), you don't have to worry about vanity metrics.
When you do pick KPIs, ensure that you pick KPIs to tell both sides of the picture. Passive Consumption and A.ctive Engagement. If you do, depending on your business, you'll end up with two or max three. If you find you have five or seven, you are not doing it right.
Avinash.
The bit about creative agencies truly resonated with me. They are all silos. As a consultant I've tried to hold the mirror to their reality but it is a little incredible how hard they work to not see the reflection.
Your AI posts point out how the tactical and some strategic elements are solved by machine learning, that might put the kind of pressure needed in the ecosystem to solve for the global maxima. Creative agencies – others – will have to balance for not just being creative but also taking into account if they are solving for what happens beyond the ad impressions.
The marketing industrial empire made me chuckle. While an hyperbole it seems quite apt from my experience working on both sides with interactive agencies who focus on display and video advertising. With all the embedded complexity it is difficult to understand the "machine" and the "machine" puts its goals primary when compared to those of the clients.
We are currently only focused on the picture you refer to as passive consumption. A part of this is data, part the poor results we see when we try to prove business value of brand advertising, and part is the fact that the clients are happy to just spend money on brand awareness and consideration. I understand why this might be a problem in the long run. At the moment there is simply too much inertia in the machine.
All that said, your post is making the rounds at our agency and challenging us.
Thank you for sharing another great article, Avinash! Today, I work with entrepreneurs and they naturally and intuitively try to solve for the global maxima :)
When I worked developing software for a big company, it was common to have different points of view from different areas persons.
One method that helped us a lot to focus on a common vision was Kanban, specially the Kanban Board. For example, it shows clearly the "before and after my job", thus forcing us to abandon siloed thinking.
Maybe the use of a Kanban Board can help to visualize and solve for a global maxima in marketing as well.
Thank you again!
This article is awesome. It is hard to find this type of expansive thinking in the blogosphere.
Our current existence is sadly siloed per your picture. If I assess it through the filters in this post, the problem is the incentive structure. Each part of the value chain is incentivized to solve for its narrow view of the sky. We have to change this quickly.
I have an example of the silo breakdown from today. With a company as large as Dell, this stuff is still missed blatantly.
Here is a screenshot. https://s31.postimg.cc/fh9bhbxbv/Dell-_Support.png
Talk about the right hand not knowing what the left hand is doing? I bet the marketing team spent days crafting the perfect message, only to have the IT team or bizops team change one tiny constraint….The reply email.
Brad: OMG, that is heartbreaking.
Dell sadly does not have a great reputation when it comes to support. Hopefully the leadership will consider solving for a global maxima.
Avinash.
It is hard to find this type of expansive thinking in the blogosphere. Thank you for sharing it with us Avinash.
In my experience this is a easier philosophy to implement up to a certain size, after that the company's desire to keep control over people and activity takes over. After that, it all falls apart. The end result is each little unit optimizing for what's good for that little unit.
I concur that painting the two pictures for the senior management team will cause them to ask the right questions. It might cause some leaders to even think differently about how they determine success. This can be of value.
Thanks for this great thinking.
This sentence says it all "Over 90% of advertising is passive consumption. This means that the ad is in front of the human and they may see it or not see it."
Excellent article.
Thank you very much for sharing your wisdom
I am reading this looking for inspiration.
Right now a complimentary brand asked to partner with my company who works in grocery delivery to give a coupon code equal to whatever we define our CAC for our service as a reward to customers who purchase a bundle offer off their website.
They would pick up the advertising cost, handle all the creative; and offered to even structure the Facebook-serve list as anonymized to not send to current customers, just new customers.
My company is hesitant as they have never done anything like this before. How do I help them break out of their silo'd thinking and not feel threatened and recognize that the customer already has purchase intent and synergy between the two brands customer pools.
Kenneth: It is a difficult problem to solve (more so in older organizations as old mindsets are hard to break).
I try to see if there is a decision maker with budget that will allow us to run a small experiment. Say, for a month, take 40% of the Paid Search budget (or a different bucket) and use it under your new model. Proving works best with real data rather than promises (and I'm sure you are going to win).
Another strategy is to show the fully loaded costs of doing it the way that it is currently being done, vs. the new way of doing it. Incrementality established this way might allow minds to change.
All the best!
Avinash.