Blog

Find™ The Only Personalized Commerce Search Engine Using Three Layer Personalization (3LP)

Personalization is often synonymized with relevance, and relevance is delivered effectively when there is an understanding of the context of a situation. At a very granular level, context as it is related to Commerce Search is derived from understanding who the user is, what the user is communicating, and in what channel the user is communicating their message. It is the ability to understand who an individual is and the context in which they are searching. It is knowing what he or she is searching for, on what device and most importantly how successful their search was. Without this understanding, Search query results often fall flat, lack relevance and disappoint users.

To enhance the user experience, drive clicks, conversion and ultimately customer satisfaction RichRelevance has created a Commerce Search solution that considers context and incorporates three layers of personalization.

RichRelevance Find™ is an innovative new Commerce Search solution that applies personalization at indexing, query time and at the dashboard level so shoppers receive more relevant results – in real-time. Find develops comprehensive affinity models for every product subcategory, product attribute and shopper. Then Find automatically manages how much personalization is applied as shoppers interact to get the most relevant products in real time.

This cautious and finessed type of personalization can only be accomplished by incorporating various levels or layers of personalization at different points in the Search functionality. With 3LP results are highly curated and personalized when appropriate and lightly peppered with personalization when catalog browsing is deemed more suitable.
RichRelevance Find™ delivers unmatched personalization for every Commerce Search experience by deploying 3LP into every query for every shopper, anonymous or recognized.

To learn more about the Three Layers of Personalization and how they enhance the customer experience please read our brief: Find™ The Only Personalized Commerce Search Engine Using Three Layers of Personalization.

 

The Most Wonderful Time of the (Ops Team’s) Year!

While most people were getting ready to indulge for the Thanksgiving Holiday here in the U.S., the {rr} Ops Team was stocking up on their Mountain Dew supply and compiling lengthy playlists to weather the busiest shopping period of the year. As we geared up for these events, I wondered what we would see over the upcoming days given the heavy focus on ‘Cyber Week’ in lieu of in-store Thanksgiving shopping, how much would traffic and sales spike across our 240+ retailers come Black Friday and Cyber Monday? The anticipation was exhausting. We added 20+% in capacity, a new datacenter in London, England as well as Sydney, Australia. We doubled our network capacity, and so much more to prepare for these 6 days.

The {rr} Ops team tracks a number of metrics like requests per second, response times, model builds, transfer rates and system loads during the shopping rush to ensure we are on par with how retailers expect performance. Records were quickly broken and as in years past, the numbers spoke loud and clear. In the past six days, we’ve delivered more than 38.9 billion recommendations (up from our previous record of 19.5 billion), over 630 million page views on Black Friday (up from 532M last year) and managed more than 18,330 peak requests per second on our servers. In general, we saw much higher overnight (USA) troughs in our charts, showing a larger engagement overseas during Thanksgiving night and early Black Friday. In fact we saw our EU datacenters enjoying high numbers until 1130p to midnight GMT and USA shopping up to 2am PT. Through all the holiday extended weekend we enjoyed 100% uptime as well.

Here’s a quick recap of our ride as we scaled the peaks of the Devil’s Horns in 2016. My seventh at RichRelevance – climbing these peaks never become old.

November 23: T minus 1 to Thanksgiving

  • In the two-day ramp up to Black Friday, we saw a 20+% increase in requests per second (from an average of 10,000 RPS last year to over 11,975 RPS this year).
  • On Monday of Thanksgiving week, we saw over 343M page views, compared to 300M page views last year
  • Traffic doubled and ran higher and longer globally starting in early November. A normal day from September, as shown on the below left, would show a defined growth to a peak, short idle, then a drop —think of a hill or small mountain’s shape. Whereas our troughs (nights) now rested around a sustained 6000 requests globally per second, a month ago they were below 3000. Seems like those earlier, longer holiday shopping periods sure are evident online.

 

 

November 24: Turkey with a side of shopping?

  • In all the years my team has been tracking this, we typically we see higher page views on Thanksgiving compared with the rest of the weekend, but more purchases on Black Friday and Cyber Monday. Until last year. And now…
  • Today, we hovered around 12,500 to 13,800 RPS all day.  We peaked at 13,950K RSP which means, Thanksgiving 2016 had a lot more Pageviews and clicks than last year!

 

 

November 25: Black Friday is Upon Us

  • At 7:30am PT on Black Friday, we hit over 18,3000 requests per second, breaking the record of 17,240 set in 2014.  The long high flat plateaus of Thanksgiving, Black Friday and Cyber Monday signal lots of traffic. 300RPS brings more than 1M more requests in an hour – all those clicks and views add up.
  • Page views for Black Friday were 16% higher than last year at >630M.  Our former high was previous black Friday with about 532M page views. After two years of this pattern, we are converting our thoughts to Black Friday being the king of Page Views, usurping Thanksgiving from its former glory.

 

In the graph below, the purple shows current page views in millions up through the end of day black Friday. Orange is the projections through the end of the year. Each color below represents the previous 8 years (2008 to current).

 

 

The Weekend

Time for resting, restocking our Dew, and refreshing our playlists. The typical “eye of the storm” Cyber weekend occured. Saturday’s and Sunday’s high is down well below the trough of Thursday night. We just floated around 10-11K on the Load Balancers(edge) RPS all day.

November 30: Cyber Monday

  • Our high for Cyber Monday Page Views were decent at 503M (503,359,235 to be exact).
  • Monday beat Black Friday on Units (total aggregated items placed in purchased carts over the day) inferring possibly more value/’Dollars’ on the day, but lower on clicks, pageview and the rest of the stats.
  • We didn’t break our 18,300 RPS peak from Black Friday as I had thought we would, hitting just 17,140 RPS, but still up from 2015 Cyber Monday RPS of 16,530.
  • Below is the Devil’s Horns for Cyber Monday.   As I mentioned above, we hit 503M page views as we continue our upward trend year over year (499M on Cyber Monday 2015 and 215M on Cyber Monday 2014)

 

 

The shopping season is in full force, and almost a week longer this year with the early Thanksgiving, but to summarize a great holiday shopping season opening weekend, in 6 days we had:

  • 2.7B page views,
  • 4.4B placement views,
  • 38.9B recommendations served,
  • 100% uptime on all 14 of our front-end data centers,

And a RPS peak of 18,330; that’s 65M requests per hour, more than 1M a minute.

Some things just don’t change. Black Friday 2016 was a great shopping day as expected, but brought whole new twists to existing patterns. Traffic climbed earlier in the day, plateauing high across the day and ran high for longer periods; the troughs were high, higher than any past year. We saw more traffic from overseas, and not just from growth in our client base ­ it was also much longer and much higher in EU than any year past, even APAC shopped tons. Interestingly, mobile made up a much larger percentage of sales, the National Retail Federation said more than half of all smartphone and tablet owners used those devices to shop over the holiday weekend. With Black Friday 2016 over, we prepare to wash, rinse, repeat.  2017, bring it.

I do get a bit reflective this time of year. I’m thankful to be blessed with one of the best Operations Team in my career who are like family to me.  My team has four members with more than 7+ years with RichRelevance, and we have an average of four or more years with the Company. You’d think this would just be second hand for them. But these highly skilled people worked tirelessly over each of the Thanksgiving/Black Friday/Cyber Monday weekends together and remained as focused as our first time around. Just the right amount of caution, risk, nervousness and confidence ensured that we were up 100% again, for seven years and counting.

 

’Tis the Season to Optimize

It’s almost the most wonderful time of the year! And we know it’s also retail’s busiest, so we’re here to help you get in the holiday spirit and also make sure you get the most magic (and ROI) out of your personalization platform. As holiday shopping is on everyone’s mind, RichRelevance’s very own Kris Kringle, Alex Ciorapaciu, shares some tips and insights on how you can maximize on personalization (and cheer) this holiday season.

In this series of short videos Alex has personally selected some optimization tips to help set you up for success this holiday season.

Videos in this holiday series include:

Boosting Product Attributes
With days dedicated to limited time offers and promotional pricing, consider creating holiday or even ‘deal’ specific rules to ensure promoted products and categories are top of mind across your site. Alex’s step by step guide shows you how to create holiday rules within minutes right in the RichRelevance dashboard.

Customer Preference Center
With Customer Preference Center (CPC) enabled, shoppers can explicitly tell you what products, categories or brands they like and dislike and RichRelevance harnesses that information to influence the weight algorithms place on those attributes or categories. CPC helps make personalization more robust, but we never want to confuse personal purchases with gift purchases. Watch Alex’s video to uncover a quick way to eliminate this concern.
If you don’t have CPC enabled and want to learn more about it please reach out to your account manager discuss setting it up.

Holiday Strategy Messaging 
Personalization comes in many forms and in many ways what you say is as important as how you say it. When it comes to recommendation messaging customers say that when brands use their own language they’re 35% more likely to engage with a recommendation placement because it feels more genuine to the brand. Customizing your strategy messaging is easier than ever within the RichRelevance Dashboard, Alex shows you how.

King of the Hill
Ever wondered how RichRelevance algorithms and strategies really work? How they know which strategy to show to which shopper? In Alex’s quick video he explains how King of the Hill eliminates the need to worry about manually manipulating recommendations.

Dynamic Landing Pages
With RichRelevance you can create dynamic landing pages for any segment you dream up. Alex walks through how you can ensure you’re getting the most money from your holiday search engine advertising spend with custom holiday landing pages by creating rules with specific date ranges or referrer URLS.

Ready to unwrap your holiday tips? Watch the video series here.

Double Award Win for RichRelevance and Shop Direct at the Retail Systems Awards 2016

Last night RichRelevance and Shop Direct were crowned as leaders at The Retail Systems Awards at the Millennium Hotel, in Mayfair, London.

RichRelevance won ‘Online Solution of the Year’ for its personalisation project with innovative retailer, Shop Direct. The winning solution was chosen for the superior online shopping experiences created for very.co.uk customers. Shop Direct were also crowned ‘Retailer of the Year’ for their Personalisation initiatives with RichRelevance.

Retail Systems Awards

Now into their 11th year, the Retail Systems Awards recognise technology excellence and innovation within the retail sector. Winning entries come from retailers and technology suppliers who are leading the way in areas such as mobile, online, payments, loyalty and personalisation.

The awards ceremony was attended by retail and technology industry leaders from across the UK. The Retail Systems’ judging panel included Clare Joel, Head of IT, French Connection Group, Nadine Sharara, E-Commerce & CRM Director, Moët Hennessy, Sharon Peters, Programme Manager International Supply Chain, Marks and Spencer and Michelle Stevens, Editor, Retail Systems.

Congratulations to all the other winners of the Retail Systems Awards. A great night was had by all and we’re proud to have been among other industry leaders from across the UK at the awards ceremony in Mayfair, London.

 

Global Consumer Attitudes Towards In-Store Technology

RichRelevance research unveils consumer attitudes towards in-store technology are markedly similar across America, Britain and Europe, except when it comes to facial recognition.

Whilst American and British shoppers welcomed a lot of shopping experiences they were the most creeped out by personalization initiatives involving facial recognition. Out of all the countries surveyed, the French were identified as the most open to in-store technology, with the highest cool ratings against questions.

As depicted in the infographic, all the nations surveyed agree that being able to scan products on their mobile device in-store to read product reviews is the coolest trend. 79% of Americans, 76% of French, 73% of Germans and 62% of British, all think this capability is ‘cool’.

However, differences appear when facial recognition is concerned with the French being the only nation surveyed to think it’s cool to be identified as a high value shopper via facial recognition technologies (62%) verses all the other nations surveyed in agreement it’s quite creepy – US 67%, UK 75% and Germany 43% creepy.

While 40% of British and Americans are keen on digital screens in dressing rooms, the survey found the French and Germans are even more enthusiastic with over 65% of French and 61% of Germans thinking it’s cool. Likewise interactive mirrors which model outfits were considered 42% cool in the UK, verses 63% cool in France and 58% cool for Germany.

Even where shoppers were creeped out by certain initiatives in-store, the French were less ‘creeped’ out than their British, American and German counterparts. For example where 75% of British, 64% of American and 48% of Germans would not like to be greeted by a sales person who identify them via their mobile phone or app as they enter the store, only 36% of French are turned off by this idea.

It’ll be interesting to see how attitudes change in the forthcoming year in particular with virtual reality technologies becoming more widely adopted by retailers to enhance shopping experiences in-store.

Find out more about the Creepy/Cool survey here.

RichRelevance - Kicking Off Holiday Shopping Infographic

DOWNLOAD PDF

Decoding Your Recommendations Performance

Product recommendations, also known as “recs,” are a cornerstone to an effective ecommerce merchandising strategy. When fully optimized, recs typically increase retailer revenues by up to 5%. That’s a substantial contribution when extrapolated across annual sales in the hundreds of millions or even billions of dollars. However, one thing I’ve noticed while navigating the world of ecommerce personalization is that the word “performance” is often misused in the context of recs. The reality is that there are multiple meanings to this word, and for each meaning, there is a discrete method for assessing its relationship with your shoppers and your revenue. So, let’s decode your recs performance.

At RichRelevance, we group performance into three analytical categories: value, engagement, and cohort analyses. While each type is useful, each tells us vastly different things. Misappropriating the type, like confusing an engagement analysis for a value analysis, can lead to disastrous business decisions. I’m here to protect you from making such mistakes.

First, let me summarize the three types of analysis:

  1. Value Analysis: this indicates the incremental revenue impact of recs, and is usually the report of most interest to retailers. The key metrics are typically revenue per session (RPS), conversion, average order value (AOV), and the lift contribution is determined by running an A/B or multivariate test that compares key performance indicators (KPIs) of populations based on their exposure to recs.
  2. Engagement Analysis: this quantifies shopper utilization of recs, and can be a signal for relevance. The key metrics are recs sales and clickthrough rate (CTR). Recs sales represents the sale of items clicked on in recommendations.
  3. Cohort Analysis: this profiles the spend patterns of shoppers that choose to use recommendations vs. those that do not—comparing the RPS, conversion, and AOV of shoppers based on their engagement with recs.

Most retailers stumble with using engagement or cohort analyses interchangeably as indicators of value. I get it; these reports are readily available, whereas conducting a true value analysis requires an A/B test which can take weeks and expose the retailer to undue opportunity costs. However, the reality is that these reports don’t convey the incremental benefit of recommendations, and here’s why:

Disassociating Engagement and Value Analyses
An engagement analysis tells us how much shoppers use recs and, to some extent, indicates relevance. If the recommendations are random, shoppers won’t click or buy from the retailer. That said, striving to maximize recs sales or CTR is equal to saying that recommendations are the most important piece of content on your site—and we all know that they’re not. What’s most important is getting shoppers to convert with the highest level of spend. Recs merely support that objective.

In fact, certain kinds of recs engagement can have a neutral or negative impact on your business. It’s the responsibility of the recommendation technology and how it’s configured to mitigate these instances. To shed more light on the matter, here are three events you should be aware of that explain why value is not a function of recs engagement or sales:

  • Overlap: Sale of items that would have happened anyway, even in the absence of recommendations. Frankly, the vast majority of sales that happen through recommendations aren’t incremental, so indiscriminately increasing recs sales does not guarantee more cash in your coffers.
  • Cannibalization: Reduction in AOV due to the presence of recs. If recommendations cause a shopper to buy a cheaper SKU than they otherwise would have purchased, that takes money out your pocket as a retailer.
  • Shopper Distraction: Reduction in conversion due to recs. If recommendations are optimized to attract as many clicks as possible, they can bait shoppers to click in perpetuity until they become fatigued and leave your site without converting. No one wants this.

And of course, we have copious amounts of data to demonstrate that maximizing recs engagement has a detrimental impact on revenue. We’ve plotted recs CTR and recs sales against RPS lift for a multitude of RichRelevance A/B tests, and the resulting scatterplot shows no positive correlation between engagement and incremental revenue. In fact, the data suggests that at extreme levels of recs engagement, RPS lift can be severely compromised. For more details, view this interesting TED Talk on the “paradox of choice”, which implies that oversaturating a shopping experience with product options and baiting shoppers into excessive exploration can result in non-conversion.

Quite simply, it’s not about maximizing recs engagement; it’s about driving the right level of engagement that minimizes the aforementioned events and maximizes your per session revenue.

Disassociating Cohort and Value Analyses
A cohort analysis tells us the RPS, conversion and AOV of shoppers that choose to use recs versus those that do not. Most often the shoppers that choose to use recommendations have substantially higher key performance indicators (KPIs), which can lead one to presume an extreme level of benefit from recs. That said, please consider that although recommendations do generate substantial revenue lift when properly optimized, a cohort analysis does very little to quantify that impact.

In the business-to-consumer world, shoppers that utilize recommendations tend to have higher spend propensities, so it’s not uncommon for recs users to have 2x the RPS of non-recs users. Intuitively, this makes sense. When dealing with household consumers, willingness to engage with your catalog is a primer for purchase intent. In the business-to-business (B2B) world, the inverse is true. B2B buyers have predictable requisition lists that they rarely deviate from. Consider an office manager that’s restocking his company’s office supplies or an IT manager that’s purchasing specific laptop models for her business. These professionals know exactly what they want, order in bulk, and do not use recommendations.

So, these non-recs users have high RPS, conversion, and AOV. Conversely, the recs users represent those shoppers that are less committed—those that may or may not know what they want, and if they do convert, it’s with much less spend. For sites like Staples, Office Depot, or Dell, these are most usually your household consumers, and a cohort analysis would report much lower KPIs all around. Therefore, due to a self-selection bias, whether a cohort analysis shows higher or lower KPIs for recs users is independent of the value the recs technology actually delivers. Rather, it’s indicative of the types of users that decide to use recommendations.

Are recs impacting your best shoppers? Do recs help monetize individuals least likely to convert, or do they have broad resonance across all levels of spend in your shopper base?

So, there you have it—your performance analyses decoded to help you learn a bunch more about your recs and answer these questions: Who is using recs (Cohort Analysis)? How much are they using it (Engagement Analysis)? What incremental value does this utilization provide (Value Analysis)?

1 2 3 32