Get fresh updates from Hortonworks by email

Once a month, receive latest insights, trends, analytics information and knowledge of Big Data.


Sign up for the Developers Newsletter

Once a month, receive latest insights, trends, analytics information and knowledge of Big Data.


Get Started


Ready to Get Started?

Download sandbox

How can we help you?

* I understand I can unsubscribe at any time. I also acknowledge the additional information found in Hortonworks Privacy Policy.
closeClose button
September 26, 2016
prev slideNext slide

Want Fries With That? Build A Digital Recommendation Engine That Generates Revenue


In the US fast food industry, this is a common question when you order a burger.  ‘You want fries with that?’   It’s in the American psyche at this point, and has become common parlance. I was recently heard this exchange:

‘Hey, can I get a copy of your targeted promos report?’   

‘Sure!  You want fries with that?’

Everyone laughed and the report changed hands.  No fries to be seen. This is one of the all-time great product upsells in fast food. This is probably the best single way to increase the purchase value.

Retail Is Struggling With Targeted Promotions

In reality, target promotions are something retailers often struggle with. Especially in the digital world of web and apps.  When a consumer is making a purchase, everything is there for upsell to happen; the desire to own, ability to buy, and availability of related products. But it’s only at that exact moment that the opportunity exists to upsell.

Herein lies the problem. If you browse most websites today, you are constantly shown offers, promotions, and recommendations. But how many of them truly offer value? For example, I was recently browsing for my car, and then later my wife browsed about quilting. Over the next few days, I was hounded by a confusing blend of car parts and knitting needles! The algorithms actually detracted from my browsing experience.

The Opportunity To Improve

Amazon has mastered the art of upsell and recommendation.  A 1998 whitepaper ‘Item to Item Collaborative Filtering’ first laid out a concept that has since grown to dominate their logic for the web, product positioning and consumer interactions. If you’ve ever visited Amazon to purchase a book and been shown what others read next to your selection, you’ve experienced it.

Here’s how powerful that is.  According to a Forrester article in 2012, nearly 60% of Amazon’s revenues are a result of this recommendation logic!  Stop for a second and think about what that number is.  And all from ‘you want fries with that?”

All digital retailers can learn from this. Yet many large retailers I’ve met with report low revenue percentages from their digital group – sometimes as low as three percent.

In fact, it’s not common to find high revenue percentages at multi-channel retailers. This is partially because of “webrooming” which is the practice of browsing online, but then purchasing in the store. But it’s also the poor record of legacy bricks and mortar retailers in truly embracing the digital space.

Why is this?  Partly because many organizations started their web presence by simply standing up a digital ‘brochure-ware’ site and calling it ‘Store 999’ in the merchandising system.

Regardless of how a digital property began, it needs constant cultivation. Without full attention to presentation, commerce, and recommendation on the web, you are missing out. Then someone else is capturing that sale elsewhere.  And even if you do capture it, are you then also upselling by expanding the basket?

Let’s Talk About Data & Personalized Recommendations

How can we fix this?  Start by focusing on improving the digital face of your brand then also ask how can we expand your revenue by maximizing the sale.

There are two main places to upsell.  During the path to purchase and at check-out. Both require personalized and relevant product placement. The consumer won’t buy knitting needles if they are looking for car parts. They won’t buy high-priced items if they are in low priced frame of mind. Here’s just a few ways data helps:

  • It’s always best to know purchase history. There are effective ways to sniff out identity, acknowledge past purchases, and recommend something new.
  • What’s in the basket? In a full basket, don’t recommend a more expensive version of the same thing! This may damage goodwill.
  • Where have they browsed? if they dwelled in sportswear, and then added some dress shoes, it’s appropriate to go back to past browsed items.
  • What have they reviewed? Chances are if they reviewed or commented on an item. Let’s look at companion items.

What Leaders vs. Laggards Do: Recommendation Engines

When managing a digital site, retailers adjust their commitment based on their goals. Is it commerce or brand? Is there a social component? Or is it more of an experience?  According to a recent analysis by RSR, in 2015, ‘only 23% of Retail Winners said the primary role of digital is to transact’ which shows the challenge retailers face to truly adopt the value of digital.


Regardless of the perceived purpose of your digital channel there are a few things you can do.  Take the cue from leading digital commerce sites and become more collaborative, design with the consumer in mind, make transactions effortless and recommend suitable products in a graceful and appropriate way. Almost as if a real store associate was helping out.  Most important, let data drive that experience.  It’s not difficult to achieve this by leveraging data if you know how.  We know how.

This leads to my recommendation.  Use all data to drive your recommendation engine.  One of our customers who did this reported an immediate 15% growth in revenue. And because they were able retire an existing tool, they were nearly a cost-neutral project with significant business impact.

Data Wouldn’t Recommend Onion Rings with a Silk Blouse

One big problem to watch out for is — do you know your customer digitally?  This is also a data question. Many retailers are losing visibility to consumer behavior. What once was as simple as a question on the store floor can now seem impossible to decode digitally.  Plus, strategies that worked in the stores will not work online.

This is mainly because the data exists in multiple formats. It’s data at rest on phones and apps, and data in motion from websites and databases. The challenge is acquiring, storing and then integrating the data.  You need the complete view but it can be obscured if you’re missing a crucial data pieces.

That’s where Big Data and things like Hortonworks Connected Data Platform come in.  We know how to make integration happen. We do this for retail leaders globally.


Retailers today need to maximize consumer interaction, which is possible through a well-thought out recommendation engine that can drive revenue and loyalty through larger baskets and targeted promotions.

Come and meet with us at Shop.Org Digital Summit at booth #5032 I’d love to talk with you about customer successes, best practices, and lessons learned pursuing a comprehensive recommendation engine.  Whether knitting needles, car parts, or fast food — make sure you are upselling the right way! And be sure to get the fries with that.

Leave a Reply

Your email address will not be published. Required fields are marked *