Which personalization techniques work best?
Photo: @dabyki via Twenty20

According to Sailthru’s “2022 Retail Personalization Index,” the three most prevalent personalization tactics are predictive personalization, email send-time optimization and buyer journey-based personalization.

Those techniques are described as:

  • Predictive personalization (cited by 50 percent of brands and retailers in Sailthru’s survey) — the ability to predict preferences and actions of users based on their past online behaviors and then customize current and future communications via the right channel at the right time.
  • Email send-time optimization (47 percent) — a feature that enables marketers to send email messages at the most optimal time to each recipient, based on when they are most likely to engage.
  • Buyer journey-based personalization (44 percent) — a tactic that caters to individual buyers’ entire shopping journey, from discovery to point of purchase.

Other popular personalized techniques used by the majority of larger retailers and brands surveyed in the study include channel-based personalization, personalized online landing pages, device-based personalization and click-based personalization.

The study found 71 percent of brands and retailers think they excel in marketing personalization, but that only 34 percent of consumers think the same.

A 2019 study from The Relevancy Group Research sponsored by Liveclicker based on a survey of 147 retailers found most retail marketers “leveraging relatively rudimentary personalization tactics.” The top-five tactics were: first name, used by 86 percent of retailers surveyed; followed by subject line, 66 percent; email body content, 62 percent; offers based on past purchase, 55 percent; and offers based on location/geography, 53 percent.

Relevancy Group wrote in the study, “While many marketers certainly realize some value from these tactics, far fewer are utilizing more sophisticated efforts such as dynamic pricing based on inventory levels, real-time data influence, and rules-based personalization that have the potential to drive more significant business impacts.”

A 2019 study from Kibo, formerly Monetate, found purchase history to be the data most used by retailers to support personalization efforts, cited by 94 percent. That was followed by email activity, 88 percent; website behavioral data, 72 percent; mobile actions, 67 percent; POS system, 42 percent; third-party data, 38 percent; and sales associate data, 37 percent.

BrainTrust

“Bad personalization is worse than no personalization at all. “

Cathy Hotka

Principal, Cathy Hotka & Associates


“An essential challenge with personalization is 360 degree data.”

Ananda Chakravarty

Vice President, Research at IDC


“The big takeaway might be that consumers/shoppers don’t think much about these efforts and marketers need to wake up to that fact.”

Dr. Stephen Needel

Managing Partner, Advanced Simulations

Discussion Questions

DISCUSSION QUESTIONS: What do you think are the most effective, most overrated and most underrated personalization techniques and related data-mining approaches? How do you see personalization tactics evolving over the next few years?

Poll

Which of the following three personalization tactics is most effective?

View Results

Loading ... Loading …

Leave a Reply

17 responses to “Which personalization techniques work best?”

  1. Liza Amlani Avatar
    Liza Amlani

    Personalization, especially via email marketing, can be so hit or miss. Once you get it wrong, the customer will remember and never trust you.

    I’ve had many emails sent to me after I have bought a baby gift where the retailer targeted me with product suggestions of everything I would need after my baby was born. I clearly bought off a registry from their website. It can be annoying and frustrating.

    1. Gary Sankary Avatar
      Gary Sankary

      I’m with you. Another factor that marketers underestimate, in my opinion, is the creep factor. I don’t like getting unsolicited offers about items based on predicted behavior. When I see those I find myself making a beeline to the “unsubscribe” button.

  2. Melissa Minkow Avatar
    Melissa Minkow

    I worry about over-personalization. Predictive personalization doesn’t allow much room for nuance, and every buying scenario varies, even for the same shopper buying a repeat item. I do think that seeing more dynamic pricing and personalized landing pages would be valuable, but I think the key is really providing all the possible options for shoppers to customize their own journey as needed. For example, allow shoppers to build different filters dependent on their preferences and situational needs, offer all forms of fulfillment to choose from in order to accommodate changing delivery requirements, etc. Where personalization hasn’t yet advanced (but needs to) is in acknowledging how different each path to purchase can be.

  3. Dr. Stephen Needel Avatar
    Dr. Stephen Needel

    The big takeaway might be that consumers/shoppers don’t think much about these efforts and marketers need to wake up to that fact. The state of the art is not nearly advanced as marketers seem to think. The more sophisticated efforts may be the ones that are more fruitful, but the price tag might be too high at this point in time.

  4. Gary Sankary Avatar
    Gary Sankary

    There’s a very fine line here between privacy and personalization. Marketers need to be very careful about how they exploit this data today. Aside from alienating a certain number of customers who don’t like being tracked, these activities are being closely monitored by advocacy groups who have had success in Europe and California getting legislation enacted to put limits on what marketers can do. I see this trend getting stronger.

    The best approach in my opinion is to soft pedal the marketing and try to develop a more consultative and engaging relationship with the consumer based on relevant content and value beyond the traditional “see a behavior – send an offer” schemes.

  5. Carol Spieckerman Avatar
    Carol Spieckerman

    Marketing is but one aspect of personalization (and the trickiest to execute effectively). All too often customer service is left out of the personalization framework. Virtual interactions offer a valuable opportunity to interact with customers and engender long-term loyalty, yet these exchanges are still generally viewed as a necessary evil. I explored the concept of predictive customer care in my latest podcast interview. Arming customer service representatives with personalization tools allows retailers and brands to personalize interactions when it matters most – during direct interactions with customers as they seek answers.

  6. Cathy Hotka Avatar
    Cathy Hotka

    Bad personalization is worse than no personalization at all. And the methods that are described here don’t address the most important personalization rule of all: ask the customer what she wants.

  7. Paula Rosenblum Avatar
    Paula Rosenblum

    I think email “send time optimization” is a non sequitur. And means not much. How do you know when the shopper is going to read it, exactly?

    I think the best personalization would be personalized home pages, but retailers must figure out the best way to get that information. Most of the time when I shop online, I feel like it’s just like a department store — organized by category, and then you can go browse. How boring!

    I think there’s a real tech business in offering the consumer the opportunity to state her preferences for a fee, and then selling that info to a retailer if the customer agrees, so that her choices can be made clear from the jump.

    I find the caveat that “if you don’t tell us what you like, we might show you irrelevant information” to be a real turnoff, and certainly not an incentive. The incentive is either a small discount, or a token of some sort.

  8. Ken Wyker Avatar
    Ken Wyker

    The problem with most personalization is the mindset of using the data to sell something as opposed to serving the customer. Customers don’t want companies to market to them better, they want them to leverage what they know to make it easy for them to get what they want. Adapting the timing of sending emails may feel like personalization because it is using customer response data, but is it really making anything better for the customer?

    The time to be customer-centric is when designing what personalization should look like. What would YOU genuinely like to see a brand or retailer do to make your experience better or make it easier for you to buy from them? When you adopt the approach that a customer’s purchase and browsing history can be leveraged to serve the customer, it opens up a whole new set of ways to leverage your data. It’s also the kind of personalization that connects with customers and drives sales and loyalty.

  9. David Spear Avatar
    David Spear

    How many times do we see surveys say that corporations believe their marketing and personalization efforts are great, but yet consumers say something very different? There is a massive gap that exists. Truth is, most retailers aren’t good at personalization. Those that are proficient, have invested in the right resources, skill sets and technologies that have enabled them to climb the maturity curve of data and analytics. Kibo’s study points to a few data elements that most retailers use (like purchase history), but I would submit that all of those data elements are important. The more the better. What about including ad search and ad promo data from Google analytics? Developing complex, high yielding algorithms is an art + science and this is where marketers intersect with data scientists to concoct the right mixology.

  10. Ryan Mathews Avatar
    Ryan Mathews

    Based on my personal market-of-one experience most personalization is overrated. That said, one day it will not just be an effective tool, but likely the dominant one. But personalization will only come into its own when we see real advances in high-level AI-driven affective computing and emotionally-intelligent AI, and we are still quite a ways away from that.

  11. Howard Davidson Avatar
    Howard Davidson

    As we all recognize, it’s a delicate balance between overstepping and making an appreciated connection. As partner in a company that uses AI technology to create personalized videos at scale, we’ve found incredible response from emails and SMS text in terms of opens and CTAs, because the personality in the embedded video recognizes the individual in a couple of relevant ways: name, location, perhaps a product type preference or anything else that’s meaningful to engagement or the recipient has expressed interest about. So perhaps it’s not just the data but is also the way the outreach or response is delivered?

  12. Ananda Chakravarty Avatar
    Ananda Chakravarty

    An essential challenge with personalization is 360 degree data. Even for predictive tools, data that’s not shared but may be highly relevant are ignored – causing consternation for the consumer.

    For example, Sue is looking to buy a new sweater, she clicks on various search engines and spots one that she likes. She heads over to the retail store and purchases it. Retargeting tools push more sweater options in all of her ads, which becomes annoying. The store purchase (offline) wasn’t registered in the predictive analytics. Even with AI tools, Sue’s lack of desire for a sweater is not captured – because it might not be a digital input. Expanding this further, even digital data might not be shared across different retailers and online sites. Until AI tools can predict timing accurately (which is being worked on) there will be a continued challenge in predictive personalization.

  13. John Karolefski Avatar
    John Karolefski

    A number of shoppers do not want to be identified through personalization. They prefer to remain anonymous either because of privacy concerns or they just do not want to be bothered. This number is higher than most marketers and retailers think. They’ll have to factor this into their programs.

  14. Doug Garnett Avatar
    Doug Garnett

    I’m certainly not an expert in personalization, but these are no surprise that they’re the top. What I’d like to see is in-depth evidence that they work. Not even Amazon can send me something I’m interested in buying despite holding 25 years of purchase information and doing extensive online tracking.

    What retailers should consider is how many times lately we’ve seen that returning to traditional targeting methods and simple personalization (including the customer name in the email) is just as effective.

  15. Dan Frechtling Avatar
    Dan Frechtling

    Over the next few years, personalization will rely on first party data — provided by the consumer to the same entity that targets them. Third party or “shared” data will require consumer consent by law. The public policy shift initiated by EU GDPR in 2018 and emulated by California’s CCPA in 2020 will grow to 75% of the world’s population next year, according to Gartner.

    It’s just as important that consumers DON’T see certain ads as it is that they do. Marketing campaigns will evolve to be privacy-first, just like ad campaigns changed to mobile-first 10 years ago.

  16. Roland Gossage Avatar
    Roland Gossage

    Today, any personalization technique that doesn’t uniquely target the individual is overrated. True personalization is one-to-one, not segmentation. Next-gen AI-powered solutions have transformed retailers’ ability to move beyond segmentation and grouping to deliver a truly customized experience across all marketing channels. By leveraging AI, the latest technology analyzes data – industry, consumer and inventory – and updates in real-time. Search, browse and recommendations can all be tailored to a customer’s preferences – such as brand, style and size – at scale, immediately, while also being ordered to maximize revenue. It’s a seamless experience that delivers value without seeming “creepy.” As machine learning and generative AI evolve, we’ll see even more personalized capabilities that can deliver true one-to-one experiences in the coming years.

17 Comments
oldest
newest
Liza Amlani
Liza Amlani
1 year ago

Personalization, especially via email marketing, can be so hit or miss. Once you get it wrong, the customer will remember and never trust you.

I’ve had many emails sent to me after I have bought a baby gift where the retailer targeted me with product suggestions of everything I would need after my baby was born. I clearly bought off a registry from their website. It can be annoying and frustrating.

Gary Sankary
Gary Sankary
  Liza Amlani
1 year ago

I’m with you. Another factor that marketers underestimate, in my opinion, is the creep factor. I don’t like getting unsolicited offers about items based on predicted behavior. When I see those I find myself making a beeline to the “unsubscribe” button.

Melissa Minkow
Melissa Minkow
1 year ago

I worry about over-personalization. Predictive personalization doesn’t allow much room for nuance, and every buying scenario varies, even for the same shopper buying a repeat item. I do think that seeing more dynamic pricing and personalized landing pages would be valuable, but I think the key is really providing all the possible options for shoppers to customize their own journey as needed. For example, allow shoppers to build different filters dependent on their preferences and situational needs, offer all forms of fulfillment to choose from in order to accommodate changing delivery requirements, etc. Where personalization hasn’t yet advanced (but needs to) is in acknowledging how different each path to purchase can be.

Dr. Stephen Needel
Dr. Stephen Needel
1 year ago

The big takeaway might be that consumers/shoppers don’t think much about these efforts and marketers need to wake up to that fact. The state of the art is not nearly advanced as marketers seem to think. The more sophisticated efforts may be the ones that are more fruitful, but the price tag might be too high at this point in time.

Gary Sankary
Gary Sankary
1 year ago

There’s a very fine line here between privacy and personalization. Marketers need to be very careful about how they exploit this data today. Aside from alienating a certain number of customers who don’t like being tracked, these activities are being closely monitored by advocacy groups who have had success in Europe and California getting legislation enacted to put limits on what marketers can do. I see this trend getting stronger.

The best approach in my opinion is to soft pedal the marketing and try to develop a more consultative and engaging relationship with the consumer based on relevant content and value beyond the traditional “see a behavior – send an offer” schemes.

Carol Spieckerman
Carol Spieckerman
1 year ago

Marketing is but one aspect of personalization (and the trickiest to execute effectively). All too often customer service is left out of the personalization framework. Virtual interactions offer a valuable opportunity to interact with customers and engender long-term loyalty, yet these exchanges are still generally viewed as a necessary evil. I explored the concept of predictive customer care in my latest podcast interview. Arming customer service representatives with personalization tools allows retailers and brands to personalize interactions when it matters most – during direct interactions with customers as they seek answers.

Cathy Hotka
Cathy Hotka
1 year ago

Bad personalization is worse than no personalization at all. And the methods that are described here don’t address the most important personalization rule of all: ask the customer what she wants.

Paula Rosenblum
Paula Rosenblum
1 year ago

I think email “send time optimization” is a non sequitur. And means not much. How do you know when the shopper is going to read it, exactly?

I think the best personalization would be personalized home pages, but retailers must figure out the best way to get that information. Most of the time when I shop online, I feel like it’s just like a department store — organized by category, and then you can go browse. How boring!

I think there’s a real tech business in offering the consumer the opportunity to state her preferences for a fee, and then selling that info to a retailer if the customer agrees, so that her choices can be made clear from the jump.

I find the caveat that “if you don’t tell us what you like, we might show you irrelevant information” to be a real turnoff, and certainly not an incentive. The incentive is either a small discount, or a token of some sort.

Ken Wyker
Ken Wyker
1 year ago

The problem with most personalization is the mindset of using the data to sell something as opposed to serving the customer. Customers don’t want companies to market to them better, they want them to leverage what they know to make it easy for them to get what they want. Adapting the timing of sending emails may feel like personalization because it is using customer response data, but is it really making anything better for the customer?

The time to be customer-centric is when designing what personalization should look like. What would YOU genuinely like to see a brand or retailer do to make your experience better or make it easier for you to buy from them? When you adopt the approach that a customer’s purchase and browsing history can be leveraged to serve the customer, it opens up a whole new set of ways to leverage your data. It’s also the kind of personalization that connects with customers and drives sales and loyalty.

David Spear
David Spear
1 year ago

How many times do we see surveys say that corporations believe their marketing and personalization efforts are great, but yet consumers say something very different? There is a massive gap that exists. Truth is, most retailers aren’t good at personalization. Those that are proficient, have invested in the right resources, skill sets and technologies that have enabled them to climb the maturity curve of data and analytics. Kibo’s study points to a few data elements that most retailers use (like purchase history), but I would submit that all of those data elements are important. The more the better. What about including ad search and ad promo data from Google analytics? Developing complex, high yielding algorithms is an art + science and this is where marketers intersect with data scientists to concoct the right mixology.

Ryan Mathews
Ryan Mathews
1 year ago

Based on my personal market-of-one experience most personalization is overrated. That said, one day it will not just be an effective tool, but likely the dominant one. But personalization will only come into its own when we see real advances in high-level AI-driven affective computing and emotionally-intelligent AI, and we are still quite a ways away from that.

Howard Davidson
Howard Davidson
1 year ago

As we all recognize, it’s a delicate balance between overstepping and making an appreciated connection. As partner in a company that uses AI technology to create personalized videos at scale, we’ve found incredible response from emails and SMS text in terms of opens and CTAs, because the personality in the embedded video recognizes the individual in a couple of relevant ways: name, location, perhaps a product type preference or anything else that’s meaningful to engagement or the recipient has expressed interest about. So perhaps it’s not just the data but is also the way the outreach or response is delivered?

Ananda Chakravarty
Ananda Chakravarty
1 year ago

An essential challenge with personalization is 360 degree data. Even for predictive tools, data that’s not shared but may be highly relevant are ignored – causing consternation for the consumer.

For example, Sue is looking to buy a new sweater, she clicks on various search engines and spots one that she likes. She heads over to the retail store and purchases it. Retargeting tools push more sweater options in all of her ads, which becomes annoying. The store purchase (offline) wasn’t registered in the predictive analytics. Even with AI tools, Sue’s lack of desire for a sweater is not captured – because it might not be a digital input. Expanding this further, even digital data might not be shared across different retailers and online sites. Until AI tools can predict timing accurately (which is being worked on) there will be a continued challenge in predictive personalization.

John Karolefski
John Karolefski
1 year ago

A number of shoppers do not want to be identified through personalization. They prefer to remain anonymous either because of privacy concerns or they just do not want to be bothered. This number is higher than most marketers and retailers think. They’ll have to factor this into their programs.

Doug Garnett
Doug Garnett
1 year ago

I’m certainly not an expert in personalization, but these are no surprise that they’re the top. What I’d like to see is in-depth evidence that they work. Not even Amazon can send me something I’m interested in buying despite holding 25 years of purchase information and doing extensive online tracking.

What retailers should consider is how many times lately we’ve seen that returning to traditional targeting methods and simple personalization (including the customer name in the email) is just as effective.

Dan Frechtling
Dan Frechtling
1 year ago

Over the next few years, personalization will rely on first party data — provided by the consumer to the same entity that targets them. Third party or “shared” data will require consumer consent by law. The public policy shift initiated by EU GDPR in 2018 and emulated by California’s CCPA in 2020 will grow to 75% of the world’s population next year, according to Gartner.

It’s just as important that consumers DON’T see certain ads as it is that they do. Marketing campaigns will evolve to be privacy-first, just like ad campaigns changed to mobile-first 10 years ago.

Roland Gossage
Roland Gossage
1 month ago

Today, any personalization technique that doesn’t uniquely target the individual is overrated. True personalization is one-to-one, not segmentation. Next-gen AI-powered solutions have transformed retailers’ ability to move beyond segmentation and grouping to deliver a truly customized experience across all marketing channels. By leveraging AI, the latest technology analyzes data – industry, consumer and inventory – and updates in real-time. Search, browse and recommendations can all be tailored to a customer’s preferences – such as brand, style and size – at scale, immediately, while also being ordered to maximize revenue. It’s a seamless experience that delivers value without seeming “creepy.” As machine learning and generative AI evolve, we’ll see even more personalized capabilities that can deliver true one-to-one experiences in the coming years.