Leaning in: Marketsmith Inc. uses data science, AI to drive marketing outcomes

Marketsmith Inc., a Cedar Knolls-based boutique marketing/media agency that uses precision data analytics and artificial intelligence to help clients attain better outcomes, has had a very busy 2019, announcing several AI-enabled products to enhance its marketing offerings.

NJTechWeekly.com spoke with Carina Pologruto, chief innovation officer, about Marketsmith’s future direction and how the new products will help the company achieve its goals.

Marketsmith differentiates itself from full-service and creative agencies by its “ability to lean in against modeling and data science, automate information delivery and do it with our years of history in this business,” said Pologruto. “(That is) a key point of differentiation between us and full-service agencies, and us and data science-specific groups or modeling providers.”

The company has been in business for 20 years, she said, adding, “We’ve always focused on how to help our clients overcome obstacles and understand the best way to use data to drive outcomes like growth and profitability.

“In the past five to seven years, we’ve been focused on data ingestion, automation and cleansing, and then applying that data to predictive and prescriptive models that help both our team and our clients understand, not just where the data is and what it shows, but exactly what to do about it.

“The products are a culmination of all of our marketing intelligence, coupled with our ability to provide data science and AI and powerful visualization through the Marketsmith IQ platform.”

Retail Intelligence

The company’s newest product is its Retail Intelligence platform, introduced this summer. 

“A lot of our marketers who are driving customers to retail, whether they have their own retail stores or Big Box retail stores, collect a massive amount of data,” Pologruto said. “It can sometimes be very messy, and the data may not feel available or accessible.”

Marketsmith IQ’s automation and cleansing of data solves problems regarding availability and transparency in retail, Pologruto stated: “It’s a highly complex environment, and we are overcoming this obstacle.” 

Once the data is cleansed, Marketsmith IQ provides a synthesized understanding of trending and performance, ultimately providing clients with a deeper view of what exactly is driving their business — whether it’s a question of certain locations or the merchandise itself.

It can also predict retail outcomes “based on dozens and dozens of intricate variables that exist in the retail market and customer communications, including the impact of seasonality, promotional events and competition,” she said.

This process helps manufacturers of retail items and their marketing staffs understand the best way to spend their marketing dollars, as they’ll know with a certain degree of confidence what will happen at the retail level, Pologruto noted. They’ll also see the outcomes quickly, receiving the information in a dynamic scorecard that shows what’s working and what’s not, thus facilitating instant optimization. Further, they’ll be able to predict the impact on their bottom line. 

“Competitive activity is part of what we are understanding for our clients,” Pologruto said.

NJTechWeekly.com asked how the system works in the real world. In response, Pologruto gave an example of a client who is “driving” its merchandise to big box stores like Target and Walmart, and doesn’t own the consumer data. 

“We are ingesting data from each one of those retail stores, understanding the client’s marketing activity,” she said. “As an example, let’s say their advertising placements are in TV, radio, print, digital, paid search, even direct mail. And they want to understand how that works together to drive retail. We will bring in all their promotional activity, all their marketing activity, as well as the data from these retail locations, in order to show them: 1. The impact of their advertising; 2. What’s going on with their competitors; and 3. What their promotions or discounts, coupons or rebates have done and the impact to margin.”

Then, Marketsmith looks at how all that history comes together, to be able to say (based on what the client plans to do in the next 30, 60 or 90 days) what its output will be by retailer, based on certain promotions, for example.

Often, Marketsmith also will provide the client with a playbook containing the marketing intelligence that it has in-house, together with what the AI is indicating, to provide a prescriptive action plan for maximizing impact. Marketsmith also will recommend adjustments in the client’s promotional strategies, price point strategies, media investment decisions and even in the client’s decision-making regarding local and national targeting.

Merchandise Intelligence

The company also brought out its Merchandise Intelligence product this year. 

“This works really well with the Retail Intelligence product, but clients don’t need to have both of those,” Pologruto said.

Merchandise Intelligence allows advertisers to synthesize a lot of merchandise-related data in order to understand performance, and it provides a detailed recommendation engine for how to optimize the product portfolio, she said. For example, Pologruto noted that an online advertiser that sells various items of apparel could go deep into the footwear category, and observe what’s going on in terms of purchases, returns, cancelations and margins. The advertiser could look at its goals for the footwear category, and see what footwear represents in relation to the overall business and all the products in the portfolio.

“We have a scoring model that would say: This is what you would need to do to optimize that portfolio,” she said. “It could be that it doesn’t make sense to invest in certain merchandise at the same level, and those products could be optimized out of the existing portfolio. Or it could find that a lot of products are showing promise in terms of how customers engage with them, and that ultimately expanding in similar products would help the portfolio overall. It’s (a stock keeping unit) line item optimization of what a portfolio can provide.”

Modeling & Data Science services and Dynamic Scorecards

Earlier this year, Marketsmith launched its Modeling & Data Science services, which involve a lot of AI and modeling relating to the media mix. This product includes trade area modeling, predictive performance forecasting, consumer cluster analyses, lifetime value, analysis and predictive scoring and marketing mix modeling.

Another Marketsmith offering, Dynamic Scorecards, involves the presentation of “strong visualizations through the automation of data and the visual representation of it, which helps our clients see a single source of truth, enabling them to see trends year over year, or over a certain date range,” Pologruto noted. “It’s very customizable, depending on what a client’s specific business and campaign (key performance indicators) are.”

For instance, clients who see that business is down 3% week over week can look at paid search, direct mail or anything else they are doing in television, and be able to see quickly which channel may be off of its goals or KPIs. It’s a very dynamic analytics tool, but it has the capability to tell clients where they have to dig in.

“Looking at a Dynamic Scorecard will tell a client immediately how their business is doing, how business is doing against the goals, and how all of their initiatives in the market have done and how they are performing against their goals,” Pologruto said.