With all of the noise around artificial intelligence for marketers, how do you understand what’s really needed to succeed today and tomorrow? Join guest IDC Analyst, Gerry Murray, and Marketo’s Senior Director of Product Marketing, Ariane Lindblom, to learn how AI empowers marketers to engage smarter, grow revenue, and save budget.
View to learn how to:
- Predict the right mix of programs for the right people to meet your goals
- Personalize your campaigns at scale with confidence
- Perform with precision, driving more value for your audience
3. only consider brands
that show they
understand and care
about ‘me.’
(Wunderman)
of buying experiences
are based on how
customers feel they're
being understood.
(McKinsey)
Buyer Expectations have
RADICALLY SHIFTED
79% 70% 66%
of customers expect
their interactions
with brands to be
personalized.
(Marketo)
14. Cognitive: A multi-disciplinary world
Source: SAS Data Mining Primer course in 1998
Computational
Neuroscience
Data
Science
Statistics
Pattern
Recognition
Machine
Learning
AI
Databases
Data
Mining
Knowledge
Discovery
23. Top 10 AI Apps for Marketing
B2C B2B Signs you need it now
1 Chat bots Very high web traffic volumes and/or online support requests
2 Virtual sales rep/Email avatar
avatar
Unable or very costly to follow up on 100% of leads
3 Real time sentiment analysis
analysis
Large volumes of social traffic, powerful external influencers
4 Recommendation engines Complex buyer's journeys and content matching, delays in decision
making
5 Live event monetization Need to optimize revenue and experiences onsite
6 Cognitive commerce Product discovery, personalized merchandising
7 Cognitive content Personalized real-time response generation at scale
8 Media mix optimization
Large omni-channel advertising spend that is difficult to attribute to
performance
9 Attribution analysis
Omni-channel marketing in complex, long cycle customer acquisition
models
10 Planning/budgeting
Need to speed planning and budgeting process and allocate investments
based on contribution to performance goals.
32. 1
2
3
Key Stakeholders
Role Benefit
CMO Improved planning, budgeting, and visibility into marketing contribution to business performance
Marketing finance Improved attribution analysis
Marketing analysts Greater scale, scope and speed of analytics
Content marketer Auto-generated content for simple interactions via SMS and email. Greater visibility into content effectiveness by segments and buyers'
segments and buyers' journeys.
Web marketer Cognitive commerce will make real time offer management systems more effective. Chatbots and virtual agents will increase engagement at
increase engagement at lower cost.
Email marketer Cognitively enabled content recommendation systems that match customer data, buyer's journey patterns, and content assets will improve
content assets will improve email performance.
Social marketer Greater insight into sentiments, influencer dynamics. And automated interactions.
Ecommerce marketer Ability to discern real time insights on how web sessions are connected via content and behavioral patterns to help build communities and
build communities and identify value added interaction opportunities
Event marketer Real time sentiment analysis and on site monetization for large scale events.
Source: IDC, 2017
Hi everyone – [introduce self]
Excited to have Gerry Murray from IDC here today, an AI expert – to talk to us about the Rise of the AI-Empowered Marketer
[CHALLENGE WITH CUSTOMER ENGAGEMENT TODAY]
Let me start with a question: How many of us are touching our customers with more email, more social, more everything this year?
Obviously rhetorical when I ask in this forum, but when I meet with marketers, the universal answer seems to be, ”yes, I’m always doing more.”
But that’s created a huge challenge for us as marketers – everyone else is doing more too. And our customers are less and less engaged with our messages because they’re bombarded.
Can I take you back to undergraduate economics for a second? I like to think of this as supply and demand. We have this amazing universe of demand drivers… but the problem is the supply curve. There is a finite supply of attention that we can work with.
We have always focused on creating demand – and that’s still the case, even as our world has evolved from one to one, to one to many, to digital.
The problem is that the supply of attention is not only limited, but it may actually be shrinking:
Part of the reason for this is simply how loud every conversation has gotten. If everyone is shouting, no single voice stands out. And buyers today are marketed to constantly at virtually every moment they are awake.
Another reason is that buyers have far greater access to information than ever before, meaning they are more sophisticated and more empowered to make their own decisions. The buyer has a louder voice than the brand.
We are in an arms race for finite attention.
Because of all this, buyers expectations have dramatically changed… To get their attention, we have to do things differently now.
Customers only want to work with brands that care about them. These days, they value experiences more than price, and feel that the best brands are delivering above their expectations across the whole customer journey.
This is Engagement Economy
Where everyone and everything is connected, expectations continue to rise, while attention is pulled in a million directions.
In the Eng. Economy, the currency is attention. And we all run the risk of getting 0 attention by trying to get too much attention
The risk for marketers in the Engagement Economy is IRRELEVANCY
What’s the opposite of irrelevancy? Again, bit of a rhetorical question, but also obvious: it’s relevancy.
We have to get relevant on a personal level, and as the data I shared a second ago clearly indicates, our customers EXPECT a high degree of personalization.
Focusing as precisely as possible on value for each individual is paramount to our survival in a world where everyone is justifying “one more send” every day.
I talk to plenty of marketers who send one more email because it costs practically nothing, but it’s a blast. Maybe to the whole database.
Or maybe, you have precise demographic, firmographic segments – and again, you’re hitting send and crossing your fingers that on an individual level, you’re going to get a slice of finite attention.
That’s the one-size fits most approach.
But be relevant, we have to take the guess work out of engagement. We have to have a clear line of sight into what engages each individual in our database.
We have to pivot to segments of one.
This is how we will win.
The other side of the equation is how much work the “segments of one” concept creates for you and your teams. It might be fine if you have ten customers, but there’s a reason we rely on segmentation – scale.
In short, hand-tuning customer experiences doesn't have the speed or scale to keep pace with the scale more of us are dealing with.
Setting up 10s of millions of pre-determined customer journeys just isn’t feasible – and still doesn’t get us to that desired state of marketing to a segment of 1 and being relevant in the moment, every time. It isn’t possible -- without help.
AI is what gets us there.
But, it has to be AI built for Marketers, think Marketer + Machine. An overlay to everything you do that amplifies your ability to Listen, Learn, and Engage to a whole new level. (Like Ironman’s mask.)
3 core tenets to our point of view on what AI for marketing looks like…
It must be purpose-built for marketers (generic AI is not the answer) – because you have to understand the context for the technology in your day-to-day so you can deploy it effectively, make sure it’s being adopted, and measure KPIs
requires a centralized place for all customer data, because AI is only as good as the customer data it has to work with… and having it all centralized makes the intelligent engagement that follows more real-time, accurate and completely informed
must provide Transparency to marketers as to what inputs are driving the algorithms and Control on when to let AI work automatically so that marketers can test and trust the outcomes
All of this provides a pathway to exponential growth by letting you:
PREDICT the right programs for the right people: to achieve business goals & go from broad audiences to segment of one (input your goals, machine guides you) -> what are the programs and campaigns I should run and to which micro-segments in order to optimize my spend to achieve my goals. ( Provide insights into customer churn, loyalty, lifetime value, and ROI to Marketing using platform dashboards or interface)
PERSONALIZE with confidence: right content, right channel, right moment, all with transparency and control so that I can test and trust the experiences being delivered
PERFORM to the Nth degree: harness all your audience data, to continually auto-optimize millions of experiences with precision, to drive more value, while making your life easier
[Hand off to Gerry]
All of this is a really compelling vision, but Gerry, I want to turn it over to you now. What does the marketer need to know about AI to start toward making this vision a reality?
Here’s Ampsy. They provide social marketing services to large rock concerts and sporting events. This is a rock concert at Gillette stadium where the Patriots play. It’s basically a pop up store with 100, 000 people in it. When the show’s over or the game ends everyone leaves and the stores closes, so its critical to optimize all the merchandising opportunities in real time. As we all know, every event has a hash tag they want everyone to use to tag their social streams. But not everyone remembers or types it in correctly. So Ampsy geo-fences the whole stadium and tags all the streams coming out of the event. They run those streams through Watson to analyze personalities, sentiments and to determine connections. They can then figure out who’s having a good time, who’s not, who’s influencing others, who’s commenting on who’s tags. They use this data to then send personalized offers via text and social media to people that have high propensities to buy or influence others to do so.
Change mgmt. how to go from ad org working with agency to a business optimization team?
Data Scientist
MS in Applied Statistics, Mathematics, Econometrics, or other discipline related to Time-Series Analysis, Machine learning and Forecasting, or other related discipline with PhD strongly preferred
Deep expertise in the design and fundamental principles of statistical modeling, predictive modeling, and applying statistical models and machine learning in a corporate or product context
Experience in leading analytical product development or projects
Documented experience working with large, complex data sets, advanced data modeling and designing analytical systems
Minimum of 2 years of experience with one or more of the following statistical programing languages such as SAS, R, S+. etc* Minimum of 2 year of experience with large distributed systems like MySQL, Cassandra and Hadoop
Experience with segmentation creation, time series modeling and report creation* Experience with machine learning and working with large datasets
Intermediate to advanced proficiency with SQL, SAS, R, STATA or other high level data programming language
Basic proficiency with Python, PHP, Perl, VB, JavaScript, C++ or other programming language
Formal training or extensive applied experience with advanced statistical methods such as regression-type modeling, data-mining methods (e.g. classification trees)
Proficiency with data visualization
2+ years developing data that merges relational tables, either within a relational database or related “big data” environments (e.g. SQL Server, Hadoop)
Data Engineer
What you'll do
Continuously design, develop, and test data-driven solutions
Help drive the optimization, testing and tooling to improve data quality
Develop measurement approaches to evaluate performance continuously
Span languages as needed, depending on the data processing framework in use
Collaborate with software engineers, ML experts, and others, taking learning and leadership opportunities that will arise every single day
Work in cross functional agile teams to regularly experiment, iterate, and deliver on new product objectives
Work from our office in New York or Boston
Who you areYou know how to work with high volume heterogeneous data, preferably with distributed systems such as Hadoop
You are experienced with batch and real-time data processing frameworks like Crunch, Scalding, Storm, Kafka, or Spark
You are knowledgeable about data modeling, data access, and data storage techniques
You care about agile software processes, data-driven development, reliability, and responsible experimentation
Audiences will be able to manage ads as a streaming service the same way they manage songs
Buyers and sellers will:
Have too much information
Place a premium on timing and meaning
Use cognitive systems to broker context and intent