Submitted article by Romi Mahajan

Chief Commercial Officer, Quantarium

The world of technology is known as much for its hype as it is for its legitimate innovations.  Atone level, this is understandable. Dreamers can only accomplish big things when they dream big and followup on those dreams with supporting rhetoric. At a different level, however, it serves to dupe consumers, customers,and investors into alchemist fantasies that often defy the laws of physics.  What we need, naturally, is a balance.

With AI, and claims about it, we as a community have come to the point where we have to decide if we are okay with “fact-checking.”  Are all claims about AI fair and accurate?  Do all companies that claim to be “doing AI” pass muster in that regard? After all, haven’t many organizations referred to any and all of their “data”initiatives as AI? Have we diluted the term so as to make it meaningless?

Discerning investors have started to kick the AI tires.  Highly skilled Scientists and Engineers increasingly refuse to be beguiled by marketing claims, choosing instead to dig deeply into the code and methodologies surrounding its production before allowing themselves to be recruited by the countless organizations that seek the limited supply of talent. 

Perhaps more relevant, still, is that people in all roles ask one fundamental question:  “Is there any relevant and practical application to the AI work you are doing?”

On all these matters, Quantarium can hold its head high.  From “Real AI” to “Real Applications of AI,”Quantarium is altering and enhancing our notions of what is possible in residential real estate, the world’s single largest asset class.

Take for instance an important but mundane question, likely asked by millions of people every day.  “What is my house worth?”  Several pundits will offer several answers to this question, no doubt.  But, will the answers be accurate and in this case, what does accuracy even mean?  What data goes into the answer?  Do we have all the data we need?  Is it all available? Do the data we collect account for every single aspect of the house that could or should go into the valuation? 

Now take an extrapolation of this question with magnified scale.  Imagine you are the CEO of a large bank that “owns” half-a-million residential mortgages. What is your portfolio worth?  How much risk are you holding in the portfolio?  Are you too exposed in a particular geography or demographic?  Do you have sufficient data and the ability to process and make sense out of it?  Can you do this all with the speed that is called for by regulation and market conditions?

These questions are easy to pose but hard to answer.  Further, while these questions may not seem “sexy,” they underscore the reality of the single biggest source of economic value and for most families the single largest source of equity.  Using AI to drive accuracy, speed, and scale in this market is complex, genuine and incredibly important.Indeed, real AI applied to real industries with real outcomes is the name of the game. Therein lies the balance we seek, the convergence of both the hype and of the reality.

Marketing Darwinism caught up with Samir Saluja, Co-Founder of DeriveOne and former Microsoft Learning Executive.

Marketing Darwinism: Samir, you left Microsoft to start DeriveOne with your partner Jason Talwar. Tell us about that transition.

Samir: Microsoft played a huge role in my development and in my understanding of the needs of customers. It also helped me understand the importance of data and using data for scale. What Jason and I realized is that despite the importance of data in the organizational and business world, most were not conceiving of it correctly or in the most efficient manner. We saw that “data for data’s sake” started becoming the essential mantra for companies and we wanted to help them avoid that trap. Thus we started DeriveOne. We consider Microsoft a key partner so still are in the ecosystem.

Marketing Darwinism: “Data for data’s sake?” What do you mean?

Samir: Data has come to be seen as an important commodity and as such companies are scrambling to ingest it and even hoard it. Ingestion is great but what about digestion? How do you use this data? What about data-overload? What about useless data? What about “fake” data? The point here is that data is not an unalloyed good if not trained on decisions. In fact, we believe you need to look at the decisions first and figure out the data needed based on the decisions not just because “data is good.”

Marketing Darwinism: You have a very varied background. Tell us more about your journey.

Samir: Thanks Paul. I have spent time in the Peace Corp, as an entrepreneur, as a business-owner, as a Microsoft employee and as a volunteer. All of these experiences helped me converge on being who I am today, personally and professionally. The Peace Corp helped me learn to listen, empathize and act. Being an entrepreneur taught me about risk and reward. Microsoft helped me understand the importance scaling through partnerships. DeriveOne has helped me realize the value of focus.

Marketing Darwinism: Why should an organization hire DeriveOne?

Samir: We believe that partnering with our customers in the context of what they need and what their customers need is crucial. We help organizations, medium to large and even some startups, use data to hone and focus their customer segmentation strategy, the cognitive diversity of their teams, and to inculcate the culture of decision-driven data. We believe that methodology matters, eliciting what the true decisions that need to be made is key, and delivering accountability is necessary. We are humbled daily by interest in our company.

Marketing Darwinism: Clement, tell us about Quantarium.

Clement: The company is inspired by insights from Quantum Physics and the potential inherent in applying them to Machine Learning approaches within an A.I. framework. We have organically gathered the analytical methods of fields as far reaching as Quantitative Genetics to build a leading Artificial Intelligence company that enables competitive advantage in vertical industries via advanced predictive and propensity models along with smart decision-engines. To be sure, there remains a lot for Quantarium to accomplish and indeed we have the ambition to match, though we are quite proud of our synthesis to date and the benefits our customers are enjoying each day.

Marketing Darwinism: Can you tell us which verticals you focus on mostly?

Clement: The beauty and peril of A.I. is that it can seemingly apply to everything and that can be intoxicating, thus both market and organizational focus to execute become paramount. Our first salvo is residential Real Estate, a $20 Trillion asset class that represents in a meaningful personal perspective, the most important of all sectors because it constitutes the largest purchase a family ever makes. Quantarium looks at Real Estate from a perspective not only of Data (and there is a lot of Data!) but also of modeling scenarios of what “could” be. For both financial institutions that “own” and service mortgages and for the individuals who own homes, understanding the “deep” economics is very important. From valuations to other analytical models, Quantarium intends to revolutionize the approach and economics.

Marketing Darwinism: Clement, you have a background in large companies like Microsoft, how is it being a CEO of a start-up.

Clement: Thanks for the question. There is nothing headier than building something with world-class people who humble me every day with their vigor and intelligence. At Microsoft, I learned how to manage A+ teams and to think about products and customers at scale. Applying that to the need for speed in the startup world is my biggest challenge and joy.

Marketing Darwinism: Clement, I must ask you this. A.I. has become a “buzz phrase” …how do you distinguish yourself.

Clement: You are certainly correct about that. The technology business is very much about fashion and phraseology. Unfortunately, it is also often about false claims as well. Quantarium’s founders team, with Ph.D’s and accomplished experts in the field, undertook the approach that A.I. is best when it enhances the ability for people to both arrive at a valuable truth in a quicker and more judicious fashion, and then start to predict future truths, or certainties, given the current business exigencies. Quantarium established itself as an A.I. company from the get-go, it’s in our DNA; as a matter of fact, the first platform Quantarium built, QVM/Quantarium Valuation Model relied on M.L./A.I. technologies such as evolutionary programming when “Artificial Intelligence” was not such a buzz phrase yet. Our team consists not only of award winning Mathematicians and Engineers but also of some of the best “technology translators” in the industry. Algorithms and A.I. are indeed assets, however when you add them to the human agency and agility, you get real applications of real A.I.

Marketing Darwinism: What’s in store for Quantarium in the next phase of your growth?

Clement: Good question. While structured as a tridimensional growth approach, with a clear focus on increasing market share, innovation / differentiated I.P. and product expansion, in many ways we’ve been silent for too long. We enjoy genuine, solid relationships with our customers and partners but haven’t “splashed” in the market yet. That has been by design but the time has now come to shout from the rooftops. We’re showing up at conferences like the O’Reilly AI Summit and Strata. While we remain humble and true to ourselves, we are bold at the same time so watch out for us.

About Clement
Clement is Co-Founder and CEO of Quantarium, an Artificial Intelligence company enabling vertical industries via advanced Predictive and Analytic models, and smart decision-ing engines. As the name of the company suggests, inspired by Quantum Physics and fueled by the power and potential of Machine Learning such as synthesizing and leveraging approaches from Quantitative Genetics, towards resolving significant predictive challenges, Clement is an accomplished international professional for leveraging Data Science and A.I. as well as a proven business leader.

With degrees in Computer Science, Clement spent 14 years building large-scale and Enterprise-level software products and services in areas traversing Content Management and Enterprise Search. Responsible for strategy and product development, Clement directed enterprise teams for Microsoft such as SharePoint, and MS Enterprise Search, while building a proven track-record for recruiting and developing teams with exceptional culture. Prior to Microsoft, Clement started and ran several software companies.

Originally hailing from Romania, Clement lives in the Seattle-area with his wife and children. He is actively involved in a variety of philanthropies and applies philosophy to technology as he builds lasting companies.

Data is the watchword in organizations large and small. In fact, how an organization frames data is the single most important determination of future success or failure. As some put it, Data is the new “oil,” the commodity of most value in the modern age.

Many business leaders understand this intuitively. As business-users in the organization are forced to make larger number of critical decisions with larger “payloads” on a more frequent basis, the idea that these decisions must be data-driven is at the fore. Gut instinct is fine but gut instinct inflected with timely, contextual, and comprehensive knowledge of relevant data is a winning strategy.

While the idea of being “data-driven” is fundamental and powerful, most organizations fall short. Intentions are necessary but not sufficient. For most organizations, the technology and operational infrastructure that defines their “data” is predicated on notions that made sense in an earlier era in which there were simply less sources of data and less change to existing sources. The “size” of the data question makes for a complexity that is not pre-defined and therefore the solution to the data problem has to be flexible and adaptive. Data infrastructure maturity is necessary in today’s business environment and has 4 basic qualities: Governance, Security, Agility, and Automation.

Without these 4 qualifiers, 2 core facets of the solution are absent- democratizing access to data and liberating IT from the backlog and fatigue associated with constantly-changing business needs. Business-users work in the “NOW” timeframe while IT has its own rhythms. In order to truly be data-driven in a way that scales, organizations must empower business-users while simultaneously freeing IT to innovate. While there are cultural hurdles to this state, the biggest blockers are infrastructural.

Until very recently, good enough was, alas, good enough. The internecine conflict between Business and IT was considered just a fact of life, a “cost of doing business.” With automation technology, business users’ data needs can be managed on the fly and without the need for reactive hand-coding, conferring agility to the business teams and handing time back to the IT teams to innovate and more resources from lower value tasks to higher value tasks. This structural win-win is available today and harmonizes the needs of Business and IT.

If data is the new oil then an infrastructure to capitalize on it is necessary- an infrastructure that is mature and “Hub”-like. While all organizations are different, they are similar in their data needs and the data platforms that win will accommodate diversity and change inherently.

Guest post by:
Romi Mahajan
Chief Commercial Officer, TimeXtender

Here is a recent interview with my new friend, Gregg Thaler, is a self-professed data quality junkie and the Chief Revenue Officer of RingLead. We discussed some best practices in data. I hope you enjoy!

PD:     How often is poor data the downfall of a marketing campaign?

GT:     Well, if I said every time, that would be a bit of hyperbole, but only a bit. Typically what I find with lists – especially trade show lists – is that they are the Typhoid Mary of duplicate creation in your database. Very often if you think about trade shows, who attends your booth? Many times it’s your customers or existing prospects. And booth staff will scan them indiscriminately. If you import that list without taking preventative measures, then it creates a duplicates horror show.

So, there’s another category of bad lists and those are the lists that are purchased from vendors. And there’s a very fundamental reason why lists purchased from data vendors have data quality challenges. The challenge with contact data is that it ages like fish and not like fine wine. It gets worse as it gets older, not better. Data is foundational. CRM and marketing automation are merely vessels, it is the data they contain that is the true treasure. It’s the single most critical element when it comes to determining revenue success of failure. Everything else you do further down the order of operations, its’ outcome depends completely on the input. What do you get from an automated process when you put garbage in? Garbage out.

PD:     What would you say is one of the biggest mistakes you see B2B marketers making when it comes to data?

GT:     Well, the mistake that I’m going to describe isn’t really limited to marketers. If you want to really truly recognize the strategic benefits of having optimized data, you have to have a mindset to prevent errors at the source. There’s a simple reason for that. Generally speaking, whatever the cost is in the enterprise to get the data right at the point of creation it’s going to cost you ten times that later to fix it. But then, of course, if you do nothing, the damage caused by inferior data could be much worse, 100x worse is possible.

So often marketers come to us, and their hair is on fire they’ve got to dedupe their database right now. Yes, they’re right. You do have to remediate the situation since almost every single contact database is riddled with duplicate and non-standard data. Often, however, I hear brilliant marketers say something incredibly dumb, they’ll say “…we’ll worry about the prevention later.” Are you kidding me?

PD:     Who ultimately owns the data? It is really sales who owns it, or is it really marketing?

GT:     That’s a terrific question. Where should data governance reside? Who is the data steward? Traditionally it has been IT. Increasingly we are seeing the data steward role reside in marketing and sales which is where I rightly believe it should belong. Now, what we see in the market, most often the actual people who perform these data janitor-like tasks, are usually in marketing operations, followed very closely by their colleagues in sales ops. Really the best practice is, I believe that organizations should have a data quality center of excellence typically reporting into sales and marketing operations. Even if the COE is one person.

PD:     Let’s talk about the best way to boost productivity in marketing.

GT:     This has, of late, become a favorite topic of mine. I think it’s something that’s very under-focused on as it relates to peak performance not just for marketers, but throughout the organization. What I’m talking about is first the batch normalization of an org’s data and then the automated enforcement of data standards with technology.

I spoke to two very well-known technology companies. I asked them “how many technologies do you have in your marketing stack?” One responded 35 and another said 22. When you think about the performance efficiencies of all of those applications operating on, what for most people, is a completely non-standard set of data. That contributes to very poor application performance which would then translate into poor performance across the entire marketing stack. It’s a silent killer of revenue and application performance. People probably aren’t aware of how well those applications could perform on a standardized data set. They’ve never seen one. In the benchmarking we’ve done, we’ve seen eye-popping performance gains of up to 600% in duplicate detection in a standardized vs. non-standardized data set. Compound that kind of application performance improvement across the 35 apps that modify that data set and the gains in application performance can be off the charts.

I am completely convinced that data standardization is the single most impactful action, a unique foundation-level enabler that marketers and sales professionals can take to optimize their data for revenue performance and their applications for speed, accuracy and efficient operation.

Responsive design brings a variety of benefits – both for brand marketers and the consumers interacting with content across multiple devices. According to data from a December 2012 study conducted by eConsultancy, nearly 70 percent of client-side marketers described their experience level with responsive design as “average” or better, and more than half of that group described their companies as “ahead of the curve” or “state of the art” when it came to the design technique.

As advanced as some brands and in-house digital marketers may believe they are in responsive design, there are still quite a few challenges that optimization experts and digital marketers must heed. Here are some lessons CMOs can use to get strategic in their approach while driving real, impact-filled growth to the bottom line.

Rule #1: Don’t Become Complacent

Since the mobile marketplace is extremely dynamic and the mobile consumer is ever changing, don’t become complacent just because you launched a responsive-design site.  At the beginning of 2013, tablet users were already showing a higher conversion rate than desktop shoppers. Moreover, 20 percent of mobile users use it as their primary device. This means consumers are evolving more quickly than you may think, so closely monitor your analytics. In fact, companies like Gilt have seen a 100 percent increase in mobile users in a single year.

Rule #2: Always Be Refining

Continually develop and refine new design iterations that work seamlessly across multiple screen sizes and functionalities (e.g., touch, swipe). Being immersive is just as important as being intuitive. Pinch, swipe and zoom are features that smartphone and tablet users know and love. In the early days of responsive design, it was said that these were features that couldn’t be tapped into. With today’s more common blend of adaptive and responsive design, we know that this is not the case. Developers have touch-screen-specific controls at their disposal, and customization can be achieved through injection of JavaScript, for example. It’s important to strike the right balance between optimal performance (page-load time) and customization, as the two are interrelated.

Rule #3: Never Stop Testing and Learning

Always be testing and learning with your responsive-design site so that key information and functions are visible, prioritized and accessible to people regardless of what device they are using. Getting shoppers to move through the entire funnel – starting on the home page and moving to key product and landing pages all the way through the checkout process – is no easy feat. Each consumer, be it a first-time visitor or a longtime brand advocate, wants something different and unique from the checkout process. For a big-box apparel retailer, for example, free shipping can prove effective in rewarding high-value customers and cultivating loyalty among a brand’s average customers. This is where testing and learning play an integral role in pinpointing the optimal threshold for free shipping to boost online sales and grow the brand’s market share amid competitors. The data and insights delivered from a test-and-learn strategy could very well disprove brand assumptions and, in turn, generate the type of ROI brands seek such as higher average order value, as well as an increase in purchase conversions and overall revenue. 

Rule #4: Leverage All Data

Even the most basic site analytics can reveal huge potential opportunities.  Incorporating analytics early in the development of a responsive-design site is important. Set your responsive breakpoints you seek to track within your analytics solution and run a report for traffic to specific pages by device type. You’ll be able to glean a wealth of information about which areas of your site are seeing the heaviest tablet traffic, compared to areas with significant upticks in smartphone-only traffic.  You’ll also be able to see which areas produce low traffic or poor conversions. This tactic can help you optimize the customer experience to drive customer engagement, loyalty, conversions and revenue consistently for the long haul.

Back in 2001, when Google AdWords was just launched, I remember the day that my first pay-per-click (PPC) campaign yielded the first batch of leads for the company I was working for. In all, this tactic generated 42 leads, and a significant portion was even qualified. Better yet, the price was just right, ranging between 15 to 25 cents per click. It seemed like a great tool to grow our website traffic, as well as an effective means for generating unique leads. There was no doubt in my mind we were going to scale this campaign.

Since then, a lot has changed in the PPC world. Now there is a great deal more available in terms of competitive products from other search engines like Bing. You don’t just have an array of search ads; now there are native ads on Google that replicate the search experience, remarketing display ads, mobile ads, Facebook-sponsored ads, sponsored tweets via Twitter and LinkedIn ads. The CMO has fallen in love with performance-based ads like these PPC ad vehicles, mainly because they work (to some extent) and it’s easy to justify a budget for it when a return can be clearly shown to the CFO.

But like any ad, the efficacy of a single ad deteriorates over time because people become numb to repeated exposure to the ad. So the typical reaction is to change the ad around and run it again. But what happens when the efficacy of the ad network declines? The usual approach from marketers is to simply run more ads and spread them out in different places — all in the hopes that they will stick somewhere. But hope and guesswork do not make an effective strategy.

It’s amazing to me that so many enterprise-level CMOs focus on increasing their digital advertising budgets as the first option to increase online engagement. That’s only going to lead to a flawed strategy and less-than-stellar results. Why would you spend a lot of money, resources and staff hours on mobile ads if the online or mobile experience itself frustrates, irks and turns away customers?

What CMOs need to do is focus on creating seamless, easy-to-use-and-navigate, relevant and meaningful experiences for customers, regardless of their device or channel. And that doesn’t mean launching a full redesign of your website with fancy UX architecture, nor does it mean you should put all your mobile eggs into the responsive design basket. It means taking the slow and steady approach to test and tweak every single experience across the entire engagement funnel and using real-time data to power more personalized experiences that meet the individual needs, habits and behaviors of customers.

But before CMOs raise their hands in the air in praise of online testing and personalization, they need to make sure that they’re measuring the right metrics (that really matter for their business). For instance, analyze bounce rates and your average number of page views. These types of insights can tell you a story about your visitors — who they are, what they’re doing, where they’re going within your website or mobile site and what types of actions they’re taking. What if you have a high double-digit bounce rate of 30%, 40% or even 50%? This is what’s commonly known as the “show up and throw up” approach in the Web business. Most likely, you have a low single-digit page view of 1.xx or 2.xx, which is not uncommon. This is where the money is literally falling out of your budget. And that’s not something CMOs can or should take lightly.

Consider taking a piece of your PPC ad budget and instead, put it to good use by testing and optimizing your online experience. I’m not talking about search engine optimization (SEO). What I’m talking about is making every single page a funnel within a website or mobile site optimized and personalized for the actual traffic you are driving to it. For example, if Facebook is driving a certain portion of traffic to your website, are you using all of the data you have about those visitors and combining it with Facebook data to create the most intuitive, relevant and engaging on-site experience to convert “lookers” into purchasers? If a mobile ad is directing smartphone users to your mobile site, have you tested the specific page they’re landing on and optimizing it to drive higher engagement, conversions and cross-channel revenue? If your answer is no, then you have a serious problem.

The reality is that in a world where the consumer reigns supreme, there is an abundance of opportunities for brands to connect, interact with, speak to, engage and convert casual browsers into loyal brand advocates. So it’s high time brands stop running themselves ragged with PPC ads and start putting their attention toward creating a unified customer experience across every single device and channel.

The phenomenon of big data certainly comes with big promise. After all, having terabytes of data on customer history and behavior is certainly better than trying to extrapolate from just a few data points.

For sure, online marketers who make sense of big data are going to be better able to build customer experiences around hard data and evidence rather than on hunches and guesswork. Instead of working on intuition, or crude analytics, you could use definitive evidence to design product pages that lure your best customers directly toward the shopping cart. You’d know exactly when to introduce your promotions and offers, and you’d know which promotion would work best with each particular customer. You could optimize your online interface, so that everything from search to registration to “Place Order” was virtually friction-free.

Getting to that point, however, requires first harnessing the data. It is no small feat to integrate huge amounts of data from a variety of sources. It is even trickier to figure out exactly how to translate that information into more visits and fuller shopping carts—in real time, customer by customer.

The good news is, there are technologies and tools that make it much easier to find the gold hidden in the data—and use it to refine your online marketing with laser precision. But there’s a mind-set at work here, too—a way of thinking about data that may involve some shifts in culture, depending on where your organization is right now.

Having worked with a number of online marketers who needed to tame big data, here are six steps to help you get there:

1) Think continuous evolution and iteration, not instantaneous.

Yes, big data can fundamentally shift the way you do business. But don’t try to change everything at once. It’s far more productive to adopt a “test and learn” philosophy. Two dozen incremental improvements in site design or wording or personalization can get you much further than trying to “innovate” in one fell swoop. We see this every day.

The most successful marketers are optimizing and refining all the time. They steadily move ahead, with a thousand baby steps, finding something to improve almost every day.

Note: This tactic may call for some adjustments to Web development processes. The most agile marketers can typically go “live” with tweaks, adjustments or tests in a matter of hours. (Slow marketers wait for the next release. Don’t do that.)

2) Align big data goals with your individual business goals.

Create separate initiatives or projects for each of your business goals, such as acquiring new customers, boosting conversion rates, improving customer loyalty or increasing lifetime customer value. This approach makes it much easier to determine what type of data to reel in, and exactly how to use it. Focus a team or a project on one objective at a time.

3) Sell the concept internally.

In some organizations, moving toward data-driven, evidence-based marketing may call for some extra communication to get everyone on board:

• Encourage knowledge sharing, continual learning. Let everyone know what you found out.
• Simplify everything. Present data and outcomes in easy-to-understand terms that managers can use to make decisions. Use pictures and graphs.
• Communicate plans and achievements across the organization. Don’t hide results.

 4) Create one team for big data.

You will need to include marketing strategists, analytics gurus and Web developers. And especially creatives, who may sometimes feel threatened or hampered by having to work with hard evidence.

Then integrate with those responsible for e-commerce and site optimization. No silos allowed.

Find a committed, obsessed, dedicated executive to drive the process and act as a focus for future customer experience innovations.

5) Your own data is best. By far.

The real-time data that your website and CRM systems are gathering is far more valuable than anything you can obtain from an outside vendor. Because it’s about your own living, breathing customers, it is data that your competitors don’t have. Advantage, you.

Examples of the typical aggregate data you can capitalize on in a big data strategy:

For most marketers, the goal should be to make in-session decisions as to what customers should see, what offers you recommend and what you say to them.

Craft a custom experience for each visitor, and they’ll buy more.

Do all of this, and they’ll be back.

This article was originally published on iMedia Connection