Tough Questions for MAS Speakers
Tough Questions for MAS Speakers
We asked Marketing Analytics Summit speakers these questions, promising we would not reveal who said what. Gary Angel, Andy Crestodina, June Dershewitz, Elea Feit, Rand Fishkin, Matt Gershoff, and Tim Wilson contributed to this amalgamated perspective – a window into what Marketing Analytics Summit speakers are thinking about this year.
1. What are you excited to learn about this coming year?
2. What lesson do you continually have to repeat to your team/clients?
3. What problem are you still looking for technology to solve?
Note: this is an alphabetical list and does not correlate to the sequence of answers below – that would be telling!
1. What are you excited to learn about this coming year?
Python, Machine Learning, Quant & Qual in real time, Face recognition, Data integration, Personalization, Predictive modeling, Tableau mastery, Product/market fit, What’s next, How to help companies jump-start the basics, Balancing personalization, privacy, and regulations.
I’m hoping like crazy that I learn Python this year.
I’m excited to be part of this whole journey that digital marketing is making into the world of machine learning.
Integration of real-time analytics into the customer experience. Businesses like Madison Reed and Stitch Fix are the forefront of this and succeed by combining behavioral data with directly asking their customers questions.
Tool I want to learn: Google Optimize. Professional skill I want to learn: How to be a better manager. Personal skill I want to learn: How to listen to very small children.
Face recognition and video processing technologies.
We are really pumped to keep learning about how to best manage our customers data. Not only from a tracking perspective, but how to take data from one tool and use it in another and then do cool stuff with it. The more we can let data flow freely the easier it is to segment and personalize.
I’m excited to learn more about leveraging digital data for customer personalization on our site and in our marketing. We have many initiatives moving us further in that direction that will rely on our continuously improving & expanding dataset. Meanwhile, we are building out predictive modeling to feed back to the marketing and product teams.
How to inspire the creation of many beautiful and useful things in Tableau. Product/market fit for a brand-new software product in a previously non-existent space.The evolution of analysts’ skills, what new things are coming that we should know/learn.
I’ll admit I also find it important and interesting hearing about companies that are still struggling with the basics, because it absolutely still happens. How do you jump-start to catch up on the basics, and then quickly get to a point where you’re able to do more advanced things? (What are the organizational roadblocks…? People? Process? Tools?)
How are companies tackling and balancing ever-increasing data, technology, and personalization, while respecting user privacy and abiding by new rules and regulations?
2. What lesson do you continually have to repeat to your team/clients?
Daily reports do not have insights, Shortness of question has no correlation to effort to answer, Whys, Your opinion without evidence is biased, Signal beats noise, Remember the data could be wrong, Business goals are primary, It’s statistics, not accounting, Good data doesn’t happen by accident, Correctly specifying the problem is key, Incremental value is central, Focus.
Daily reports seldom yield “insights.” They can be a useful mechanism for monitoring aspects of the business (ideally KPIs), but please don’t think that every report will yield surprising and actionable information. “The status quo is continuing” is often the reality.
The length of your request — the ol’ “quick question” — has absolutely NO correlation to the level of effort required to fulfill it.
Analytics should always start with, “What decisions are we trying to make here?”
Be skeptical of your own personal preferences. Your POV is subjective and biased. Look for evidence to support your ideas, because every idea is just a hypothesis until it has been supported with evidence.
Filter out signal from noise, in the inbox, in social streams, in reports, everywhere.
For my team – numbers that don’t make sense are probably wrong. Even analysts routinely forget to think about whether a number actually makes sense and simply assume it’s right.
That they need to have goals and focus on achieving them. Even though this is cool, we need to make sure it impacts the business.
That data will never align 100% across multiple disparate platforms – and that’s OK! Joining across platforms always results in a small rate of error, but the data is still directionally correct & actionable. As a business, we must always be prepared to take action with imperfect data, and that includes from your digital platform. Furthermore, the integrity of that data is what you make it.
Good data doesn’t happen by accident – everyone across the organization needs to be committed to implementing & maintaining good data.
That while technology can help, inference, by its nature, is hard. So most of the value is incorrectly specifying the problem to be solved, rather than the technology used to solve it. The solution lies in the problem, if there is even is a solution.
Decision making should take place at the margin. The pertinent question for decision making is not what is the value of action X, but what is the marginal value of action X vs action Y.
When someone says, “Give me a spreadsheet with xyz in it,” the way to respond is NOT, “How soon do you need it?” but rather, “Tell me about the business problem you’re trying to solve.”
Focus.
3. What problem are you still looking for technology to solve?
Balancing personalization and privacy, Intuitive leaps, Economical tracking IRL, Technology integration, Data aggregation, Recommendations, Optimal use of human in the loop, Airport security lines.
A mechanism for providing reliable cross-channel, cross-device behavioral information for analysis without infringing on user privacy (intentionally or unintentionally, now or in the future).
Analytics still can only tell you which approach you’ve tried so far is best, but we still need people to come up with those game-changing customer experiences.
Economical tracking of people in physical spaces
Honestly, I think there is too much technology trying to solve problems right now, which is what I ultimately want it to solve. There is a tool for everything, many of those tools are half baked. I guess I am more looking for humans to help me solve all the technology problems, but this is also why my company is in business.
Easier & more intuitive data aggregation. It remains a chore to create unified datasets, whether that be uploading backend data to a frontend platform like Adobe, combining data that exists in different classifications or including costs from across a wide array of marketing partners. Making that easier will free up more time for creating value rather than creating the data itself.
Recommendations are still such a mottled, inconsistent, frustrating field. >From restaurants to hotels to home services to products — there are so few trustworthy sources, and even those change rapidly (Yelp is good in city X, not so good in city Y. Google Local is good for category A, not for category B)
The balance of “artificial” and “human” intelligence. To your question around problems, we’re looking for technology to solve – what’s that balance? How are companies striking it, between automation/personalization/machine learning, and needing to make actual human decisions and weigh in with real experience?
Airport security lines.
Get on board as a blog or media partner, share the latest news and earn complimentary tickets: contact [email protected]
To learn more about our sponsorship opportunities, contact [email protected]
We look forward to seeing you at Marketing Analytics Summit!
