Progressive Personalization: Designing a Better Personalized Experience
A framework to design for personalization, at any scale, in 2017.
Like many of you, I’ve worked both long and widely enough in digital to see personalization brave its journey from promise to pratfall to myth, and back again. Lately, it’s made the run from white whale to white-hot all over again. Personalization is re-emerging today with newfound relevance, sophistication and ubiquity.
But the past has a way of being prologue, doesn’t it?
Current status: Stranded, and stuck, in the personalization gap
In the shadow of personalization’s more celebrated successes—the Amazon chestnut that 35% of total revenues are owed to the recommendations UX—various Cassandras highlight its shortcomings: personalization subjects consumers to filter bubbles and overzealous retargeting. I regard these as speed bumps to more evolved, nuanced and satisfying user experiences to come.
The real problem with personalization is executional.
I started Bucket, a consulting and research practice, on the heels of realizing where both my client needs were going, and where my skills were soon to be needed. What I’ve seen is sobering. The practice of personalization, in short, has not kept pace with technology. Personalization remains a top priority among CIOs and CMOs, and the hallmark of countless digital products.
For many, however, the inevitable move to personalization is one of steering right for the rocks, and certain shipwreck, if user experience professionals do not step forward soon.
- It’s not just that personalization is very much a content proposition, that requires this kind of expertise at hand.
- It’s not that the solutions marketplace is so glutted and overrun, with standards and best practices few and far between.
Those are simply factors of the personalization gap as I’ve defined it.
However daunting, those factors can be addressed.
What cannot be overcome is the deafening silence among UX designers, strategists and technologists, whom need instead to explain that:
- the tooling alone is not enough, and
- what personalization needs, more than anything, to be successful is clear, rigorous, thoughtful design.
Accenture calls personalization the challenge of “[h]ow to uniquely serve everyone without overwhelming anyone,” and reflects on humans moving from 50,000 square-foot shopping experiences with 100,000 products merchandised to a 4-inch screen regaling us with 10 million.
It’s a conversation we haven’t been having because:
- personalization is a broader mosaic of various yet interdependent (or, particularly for our users, alike) technologies than is commonly understood; and
- we’ve lacked for handholds that could guide a framework to designing holistically for personalized experience.
I have set upon making a wee dent in that.
The need: A decoder ring for designing personalized experience
For all the glare of AI going mainstream—for which personalization, in its many permutations, is ever the fruit of the tree—my own consulting tour of duty in this space tells me that many seem not to recognize that solving for personalized experience—maybe more familiarly today known by its various components: testing and optimization, dynamic content targeting, bots (messaging and voice), notifications or recommendations—is already table stakes in certain domains like ecommerce, a best practice already matured into a standard expectation. At one level or another, consumer behavior indicates it will be an existential matter for yours soon enough.
Personalization doesn’t have to be a slow-motion shipwreck, both for those who shirk the voyage and those worried about where it’s taking their team. So, what is there to do in rising to this challenge? Two things:
- Navigating the personalization gap is the immediate hurdle to clear.
- After that, it’s about how to design for personalized experience. What are the “breakpoints,” in responsive parlance? What is the right mix of interactions, modules, elements or features to be personalized—and what data and information architecture will support it?
Personalization is an ambition, not a static end-point. This involves identifying and naming what personalized product opportunity means for you and your organization: a ground truth of what personalized tech can deliver to differentiate and focus the organization’s efforts, however that ambition is scoped. To see the difference up close, look at how notifications are handled as a design pattern by Pinterest and moreso an engineering function by Quora.
At the risk of self-aggrandisement, I have prepared a bumper sticker for your consideration:
Designing for personalization is about taking ‘raw’ engineering and, through a ‘cooked’ design, delivering a consistently effective, holistic experience.
(Granted, I have in mind a lengthy sticker. And a big vehicle.)
To apply this process—to cook up interactions at different scales, and in different client settings—I came up with a framework I call progressive personalization.
What’s entailed in designing for personalization?
For the wealth of literature in the wider world of personalization, search and discovery, recommendations and adjacent domains, I’ve found precious little that takes for its focus a broad-based look at designing or product managing a personalization program.
Namely, how can UXers focus and stay integrative in their execution?
Designing a program and practice of personalization deemed operationally effective and holistic is a higher bar to clear than many may realize.
For instance, this is not a domain in which “lift and shift” is ever feasible. The current hype cycle around bots and voice shows well enough the limits of merely tactical shovelware-style applications that do little based on common, typically superficial or surface-related tasks users have previously completed on sites or apps.
So how does one design coherently for a series of touchpoints that are increasingly expected to offer a consistency of experience, and carry over user data and context, in order to achieve their goals? A static design comp is fine pre-launch, but this will necessary need continual care and feeding as business rules and algorithms are adjusted.
Oh, and how does one design for a program that has many “cooks” crisscrossing functional lines of digital, IT, marketing, content and the like? (Few resources tell this tale better than John Berndt’s Personalization Mechanics.) This is part of the reason why pre-personalization, an array of preplanning activities to validate readiness to proceed with a functioning program, is a valuable early step to eventual program success.
Take a step back. Considered carefully, there is a series of preconditions to designing a successful personalization program:
- scoping a minimum viable product
- customer journey and user profiling/attribution
- a feature development focus
- data readiness
- functioning taxonomy and metadata specifications
- content strategy and content program
- benchmarking and measurement
- a broader roadmap
This list is necessary incomplete and doesn’t speak to the solutions architecture, technical buildout and implementation, or migration. It’s at about this point many will want to cognitively bow out and call it a day: this is a lot of interdependent balls in the air.
Personalization is like that.
As we plainly see, executing on either personalization program or practices is a matter, above all, of orchestration.
The outcome all these activities drive toward is simple: it’s about getting a grip on strategic and operational complexity so that personalization can be managed as a focused, integrated set of user features in your digital footprint.
Your program’s north star: A single abstraction layer
At its core, my thesis is simple.
A successful personalization effort is the progression of a single abstraction layer shared across touchpoints. If this is your guiding outcome, the rest should follow with relative ease and focus. It’s a matter of having a single source of truth about how your users, and their context, are known and how their interactions influence the UX they are presented.
This is not to proclaim that all personalization must involve budget heartburn and sprawling scopes and schedules. Far from it. It’s really about responsibly yoking the effort, objectives and impact to the ambition and resources allocated. (Another time I’ll review the virtues of t-shirt sizing one’s program as a roadmapping exercise.)
From this point, it becomes far easier to assess the sweet spot of personalization that meets those parameters, which is what I call the personalized product opportunity. From here, a designer can begin to conceptualize brand- or market-differentiating features and user stories that realize that opportunity based on the outcome (and fixed requirement) of a single abstraction layer.
If this sounds like your mission, and the associated complexity is familiar, progressive personalization could be an approach useful to you.
For our next step, we need a mental model to make good on designing out the conceptual abstraction layer. Or maybe we need a pattern library of personalization? Or a launch pad for feature and functionality ideas. How about all of the above.
Let’s go berrypicking.
Progressive personalization is a framework for designing personalized experience
Progressive personalization is a framework to help designers, strategists and technologists assess where and how to build personalization into their digital products and touchpoints, and how to evolve its use across a spectrum of technologies and applications, in a singularly focused, holistically integrated manner. While the term progressive personalization has turned up from time to time (as a swapping mechanic in 1:1 matching feature technology), to the best of my knowledge the idea has not. Today there is literature on tactical design considerations for personalized or AI settings. There are tips on how to construct a program. But there is not an established way of thinking across features, products and into the program-level about personalized experience—especially in the broad range of ways and interfaces in which personalization is delivered, from AI to recommenders, and A/B testing to machine learning, and bots to voice.
The thrust of this exercise is to provide a durable lens to measure performance, assess user value and transfer a singular, apples-to-apples way of thinking across what are often described as disparate design settings today.
Historically, when I’ve been tasked with defining a content offering, I have returned to a much-traveled four-quadrant diagram, organized along the spectrum of intent and of interactivity. In my experience, designing an experience has typically meant solving for either a lean-back, low-intent experience, for instance, or a high-intent, lean-in one—or some variation thereof. Intent is also sometimes classified as skimming and diving user behavior, and interactivity can be as simple as context and at-hand interfaces. For example, try browsing on Echo sometime, it’s painful today: voice-only interfaces are better at search, where intent tends to be higher, than a general browse navigation behavior.
Meanwhile, don’t forget that, in navigating that personalization gap, we have in the background a sense of a hierarchy of personalized experience:
- function (e.g., a search box)
- feature (e.g., notifications)
- experience (e.g., an email)
- product (e.g., an app)
If personalization can be considered not just along a spectrum of applications and interfaces but also a progression of technical and design effort, we can start to dimensionalize and assess the design choices before us.
We have in the following three dimensions an idea of three fairly distinct stakeholder interests represented:
- the hierarchy of personalized experience represents the work of digital product teams
- the dimension of intent speaks to user investment in a given interaction
- the dimension of interactivity speaks to how that interaction is designed to suit the behavior
I’ve been using these models like a trusty stool for years, but it took an unassuming conversation with Jenny Benevento to set me straight about a more sophisticated model—straight into the dojo of Marcia Bates, vanguard information scientist.
What is berrypicking
Bates pioneered the berrypicking thesis, which remains a revolutionary argument in how information environments should be designed.
As august a figure in information sciences as there is, Bates has in her career commended folks to the mission of understanding user behavior and to designing interfaces that accommodate those behaviors. Practitioners “comfortable with information, people, and technology” excel at this work—and data, users, and tech are critical legs of the personalization stool.
Her berrypicking thesis not only advanced a new modality of understanding users’ information-seeking behavior, but it also promulgated the following:
- Maximizing the different modes of information retrieval is best. “If we want to meet users’ needs, we should enable them to search in familiar ways that are effective for them,” she writes. Elsewhere: “The more different strategies searchers can use an information store, the more retrieval effectiveness and efficiency is possible.”
- One overarching architecture is nevertheless key. “A model containing a unified perspective, incorporating the full range of searcher behaviors in the information seeking process, may make it easier to design many more such features for information retrieval systems.”
- Design is the way we construct a solution. “With this broader picture of information retrieval in mind, many new design possibilities open up.”
What she gives us in berrypicking is a Rosetta Stone of fundamental, root-level interaction behavior. (Nerd alert: Not to mention how she anticipates binge watching behavior in her consideration of consuming journal runs.)
Berrypicking, decoded for 2017 personalized experience
It’s not a stretch to see how Bates’ berrypicking perfectly encapsulates some (all?) of the prevailing modes of information retrieval today as well. We see a handful of recurring modes of interaction emerge that we now commonly understand today as:
- alerts (or push media, like email)
- consumption (episodically [a single clickthrough] or serially [bingeing])
Further, these four modes snap rather neatly to the four-quadrant model:
- browse is generally high-intent, lean-back
- search is generally high-intent, lean-in
- alerts are low-intent, lean-back
- consumption is low-intent, lean-in
Considered this way, a designer’s next step is to assess their understanding of user behavior, user needs, business value and product differentiation, and then focus accordingly. (I’d like to expand on this in future.)
Seen from this vantage, berrypicking and information sciences has the potential to be a skeleton key to personalization and AI. Used effectively, this approach can open any door before you in methodically designing experiences that address the needs and opportunity of users.
Consider these behavioral vectors of browse, search, alerts and consumption as a starting point in orienting your team about its personalization regimen, and sizing what will be effective in catering to the user needs you identify for each—with a sense of the higher-order user value clearly in view (more on that below) to evaluate your progress.
How does this work practically? A client scenario
To unpack this further, let’s drop down into a client example.
A product team for an ambitious video hardware startup gave me a brief to devise features and experiences that would clearly separate them from the status quo of episode program guides from the incumbent competition, and from card stack mausoleums of the new-line streaming services like Netflix. I was armed with industry-standard data and instructed that editorialized touch-up overtop could be provisioned. I came back with some four-quadrant concepting:
- low-intent, lean-back: A suite of editorially curated “impulses” that will suggest TV and film titles prompted by user heuristics like hate-watching (it was fun specifying that from the data) and other nontraditional concepts.
- low-intent, lean-in: A matchmaking recommendation feature for children’s viewing that, querying an industry data vendor API, produces recommended titles to watch based on the age and gender of children present.
- high-intent, lean-back: A way to query “ice fishing” programming and not return results for the wider term of “fishing.” People want to pinpoint their interests, and sometimes less really is more when zeroing out noisier searches.
- high-intent, lean-in: A white-listing function that allows users to correct for recommendations “contamination” in multi-user settings (when there are no sub-profiles). In other words, a spouse’s viewing history could be suppressed or downvoted, at least at the level of recurring false positive matches suggested from the title library.
The overall idea was to differentiate the hardware’s software UX and feature set in a sustainable but remarkable enough way that it would readily surface product opportunity and marketplace differentiation. The method was to foreground the opportunity in terms of discrete user stories and lean product development, while anchoring them to a broader framework of the personalized product opportunity for the hardware client’s customers.
Progressive personalization is also a way of pinpointing precise function- or feature-level improvements that can move a personalization regime forward, without having to bet the farm all at once.
Your berrypicking may vary—that’s the point
While I have been employing this framework in my client work for a few years, I harbor no illusions that it provides perfect clarity in all situations. The argument I want to make is instead the necessity of a design architecture for personalization as a whole. Only that will maximize focus and planned outcomes, and make evolving the personalization benefit similarly successful, and the net effect greater, over time. See, for example, Colin Eagan’s model of bucketing for “tasks at hand” versus “big picture” messaging in personalized settings.
It’s perhaps also a canny way of smuggling in governance by principle and maybe instituting some conceptual headroom for organizations to be more measured and balanced in the type of messaging they send as part of a personalized experience.
Progressive personalization is effectively a single-sourcing method of coupling varied forms of personalized experience and functionality, and yoking it to a single source of truth in terms of data and content and users.
No good algorithms without architecture
Applying this thinking about the necessity of design, and about design-as-architecture that advances us beyond mere engineering, it’s similarly hard to see how an algorithm can be stood up without the supports of an architecture.
Notifications are a case in point for how success in personalization tends to reward thoughtfulness. Architecture is design; engineering is merely that. People who think notifications are simply a function and not a feature to be designed have not been paying close enough attention.
User segmentation is another one of these subject matter areas that may seem straightforward and intuitive, but in fact is a series of design choices as a profile is specified out with a series of attributes that are scored and weighed in a particular way.
The whole point of progressive personalization is making step-by-step performance evaluations
Turning our attention to assessment, there are three ways I know to consider evaluating progressive personalization.
- The first is by how you already measure product management success, as this is a program management and prioritization framework at one level. If decisions were sequenced in a way that improved delivery time or increased budget efficiency, that’s a win.
- The second is by looking to its impact on the brand, business or organization. More on that below.
- The final way I know is to look at the user value equation. How does your new feature, functionality, experience or product provide value in their eyes? Bain and Harvard Business Review have been advancing the ideas of a hierarchical model of user value, and studying its sector-specific variations. You may have user research or testing at hand to validate. (Or, like Android’s engineering team, you may have a more nuanced view of getting to user harmony; see the Appendix.)
The nice part of these last two paths to assessment is that we have tangible rubrics we can look at.
Consider experience simplification and business simplification as guiding principles that can be measured in terms of friction reduction, reduction of user flow steps or abandonment, and so forth.
Personalization should be a streamlining of experience to best outcomes for both parties, and in this way it should knit itself, at least indirectly, into the existing north star KPI for any organization. A net promoter score cannot necessarily be considered a direct outcome of a successful personalization implementation, for example, but there’s a likely link if one is controlling for other factors.
Taking a step back, there’s also an idea of brand wins, and I would commend interested folks to check out Siegel + Gale’s annual Brand Simplicity Index and MBLM’s Brand Intimacy Scale. Both are close studies of consumer sentiment towards brands around the axis of an integrated brand experience. Simplification pays in consumer sentiment, goes their argument.
Benefits & scenarios: What can I do with progressive personalization?
We’ve reviewed the means of creating a design framework for personalization, and a way of assessing its efficacy. But what might the practical point of adopting such an approach? Why take the plunge when personalization is clearly so challenging to orchestrate?
The benefits should accrue in at least two principal directions. Personalization should be about business simplification as well as a higher-impact customer journey.
- Users first: Taking user experience as your primary point of focus helps ensure the user centricity in program outcomes that so many organizations seek to achieve.
- Interoperability with growth and optimization efforts: The miniaturist approach of progressive personalization is well suited to adopting growth work and experiments in your experiences. Testing and experiments can work hand-in-glove here.
- Readiness for emerging tech: Like a broad stretched canvas, this model encourages teams to work confidently across the personalization spectrum—from recommenders to testing, from bots to widgets, and more—with a common purpose and abstraction layer, helping to minimize technical debt and maximize organizational readiness for the more robust AI tools to come.
- Risks managed: A progressive approach biases against stovepiped efforts and tactics, helping to minimize (for users) unintended friction and dissonance, and (for businesses) redundancy and risk. These are pleasant byproducts of focused execution. The ecommerce issue cited (left) shows the peril of misaligned business rules, wherein some personalized modules are accurate, and another is not—likely casting user doubt on the whole experience, when it goes noticed. This is to say nothing of the internal business alignment, and risk management, required of a personalization program, a topic explored with aplomb in Personalization Mechanics. These are nontrivial hazards of an in-flight personalization program, particularly one that is not effectively coordinated by a dedicated team.
- Guard against overreach: The most widely acknowledged domain issue with personalization is what I call “rule creep.” Principally, it’s a matter of a clumsy application of the crudest forms of personalized targeting, such as repetitive retargeting. The consumer outrage here is legion, and sometimes eloquent. Invariably, the UXers on a personalization team will be an embedded early-warning system where retargeting, or similar questionable tactics, may pose a wider brand risk.
- Learning the best tools for each task: A holistic approach takes more time than a one-tool-for-all-jobs approach, but the payoff is the knowledge and capacity it builds over time. For example, the recommenders of the world often have on-board algorithms that can be leveraged, though tuning the weights and scoring are an essential part of gaining facility with getting the most desirable outcomes from an algorithmically-driven feature or experience. Moreover, the recommender world is thick with solutions for conversion, while far fewer solutions solve for learning heuristics, or for driving content consumption. Recommenders are rather dissimilar on closer inspection. By its small-bore focus on features and functions, progressive personalization encourages a thin-slicing of the user challenge that should surface poor or suboptimal solutions among the vendor partners you may be considering.
- Prestige: Just kidding. 🤓 🤣 Personalization is moving so swiftly to mainstream practice that it’s becoming a matter of implementing it sensibly and safely: vanguard practice is more about figuring out how to leverage adjacent tech in machine learning and AI, for example. While awards and recognition are a tangible incentive, especially within sector-specific professional groups, a lot of the best practice will accrue to “fast following” on the more reliable forms of personalized experience and mechanics. Progressive personalization spreads your bets.
- Position for continual, evolving complexity: Lastly, nothing will get simpler soon. Instead, we’re headed for uncharted territory, more subsystems, and undoubtedly more complexity first. This speaks to systems integration as much as day to day operations for a personalization regimen. I’ll let these next few figures speak for themselves on this count.
A 3-stage maturity model in progressive personalization
Poise is only as good as the capability and capacity that backs it up.
I have a last thought on getting mileage from a personalization program, also tied to the matter of complexity. Impact is unevenly distributed in personalization programs, and the effects of what teams do varies on the maturity of their program. Personalized product opportunity shifts with your own advancement and capability building.
I break down maturity in 3 buckets:
- charter destinations (going places, in other words!)
This is a progression that marks out the natural and necessary learning curve of entering personalization work and I’ve yet to see significant variation.
If there’s a lesson to this view of maturity, it is to consider where you sit before lunging into a moonshot effort, particularly one that will require a series of dependencies be achieved and feature engineering to be accomplished. If there are shortcuts to launching an “all-in” personalization program, I’ve never seen them. The path is bound to be long and the journey an involved one.
Instead, why not pilot your work incrementally yet holistically?
What’s next: Designing for the future of personalization
As I’ve maintained elsewhere, I see the move to personalized technology and experience accelerating in the coming years as more and more computing becomes algorithmically driven. It will cease to be a feature of our experiences as much as permanently resident context. Today’s potent differentiation with personalized experience is really tomorrow’s table stakes.
It’s all but obvious to presume that more mature, AI-driven notifications handling will emerge as requirements of app makers from Apple and Android, for example, to which our berrypicking models will evolve and conform. That said, the core task—of speeding users to satisfying decisions and outcomes—will not.
I foresee our sophistication and finesse increasing as competition grows. If the ultimate backstop is user time and attention, the stakes of competing for that finite resource look stark—and they help explain the turn in recent years in UX study to dialing up the understanding of psychological factors, and to designing for calm, and to designing for emotion. Back in 2013, designers at Google remarked upon their usage of the work of Barbara Fredrickson, an academic celebrated for her deep study of positivity in the context of human factors like decisionmkaing. The calculus she cites, that users weigh a negative interaction with 3x the impact over a positive interaction, shows how high the stakes truly are for even the most slight hiccups in a personalized experience.
Closing the personalization gap and enhancing personalized experience in the future will be a matter of deepened awareness of human context, clearly—even as machines pick up more of the legwork. In a respect, it will be to fully appreciate Steve Krug’s classic dictum of “don’t make me think” that launched modern usability study.
Meanwhile, increasingly modest micro-interactions will carry greater weight as we’re able to design and measure them in personalized contexts. We’ll just expect more from digital experiences.
Here’s what might come next.
Today, the Yusp recommender comes loaded with nearly a dozen pre-built algorithms. Naturally, a lot of energy will be paid to more discerning and nuanced heuristics to goose discovery and engagement among users. While today that means adjusting “weights” and campaign rules or business logic, with advances in machine learning is perfectly reasonable to expect we will be able to put more sophisticated methods to use.
Design for new heuristics to solve old problems
Projecting ahead, if we believe a prevailing utility of, and value for, personalized experience is how it enables humans to make more satisfying decisions with more enjoyable or beneficial outcomes, it’s appropriate to consider what a 21st-century successor to the berrypicking paradigm looks like. Given the rising penetration of voice, bots and AI into more and more forms of everyday technology, it feels fair to hold that algorithmically-driven machines will ultimately aid and assist humans in a much wider swath of decisionmaking, from media consumption choices to how to file better tax returns.
What better way than for have AIs help us quarterback our own thinking and thought processes?
I think it looks something very much like Buster Benson’s inquiries into a taxonomy for cognitive biases. What is the end of personalization if not to speed users to satisfying outcomes—and satisfaction with the decisions they’ve made? His taxonomy is like a root-cause detector for triaging these broad situations of “too much information” or “not enough meaning.”
Benson’s side project to index and essentially reorganize Wikipedia entries provides an engrossing window into how recommender heuristics could conceivably evolve with greater computing power in the near future. Even if such an AI were only able to suggest a possible change of perspective, or highlight some of the characteristic shortcomings (biases) to looking at a subject in that way, it would take human-computer interaction and conversational interfaces a few rungs higher than where they operate today.
Much less, even if this was never brought forward to a machine intelligence, human designers would do well to consider user flows that look to this kind of heuristic deftness.
This gets at the big so-what of personalization, not far into our future. So much of the promise of personalization is bound up for me in the matter of providing discovery opportunities in a swiftly mechanizing and algorithmically driven environment.
Personalization, and the AI-driven discovery it delivers, is the Hail Mary pass for information overload
It’s easy to focus on the immediate payoff of personalization, and I want to emphasize the move to personalized technologies will also involve existential stakes. Why? Because personalization delivers on the core tenet of user centricity with machine rigor. Since we can train our machines to perform ever more supple heuristics, and deliver nuanced, highly differentiating experiences as a result, I expect we’ll find before long that personalization itself limns all notions of UX design, and of customer journeys — and goes some distance to solving longstanding, sector-specific domain issues.
Take one of my focus areas, for example. I often work in media and entertainment.
What else can rescue content providers, creators and consumers from content shock — a kind of information-overload suicide pact, also called content marketing — from the necessities of tirelessly grooming, monetizing or foraging vast libraries of content in the endless wastes of a fragmentary digital moonscape?
Naturally personalization, and content discovery abetted by AI, is rightfully hyped as the great hail-mary of content ventures the world over.
This also holds more broadly, of course, well afield of media. The so-called attention economy view is that, increasingly, only personalized or algorithmically driven messages will reach us—therefore favoring those who begin orchestrating their marketing and communications in this way. In the early days of the web, this preoccupation was findability and now it’s a matter of discoverability, and of finding users where they spend their time between, say, their inbox and social media.
Lift all the boats
I came to the personalization domain and to the Bucket effort with enthusiasm that we can do much better, right now, in how we design for personalization. Call me corny: I think there will be many winners.
It’s a matter of both professional and personal interest. I’m also working this beat because it’s not getting sufficient attention despite the outsized outcomes it offers and the growing role personalized technology will have in all our lives. To bury the lede, the stakes for ethics, culture and society couldn’t be much higher. I’m as engrossed by those challenges as I am by the business need and the sheer opportunity to ship new, more compelling modes of interactivity that drives discovery and engagement across countless communities.
UX pros: Speak up and move this conversation forward
All that’s a grandiose way of saying, I’m here if you want to talk. In its own modest way, Bucket’s role with its partner community is two-fold:
- to close the personalization gap; and
- to write a playbook for seizing opportunity through personalization.
This effort is to steward conversation, community and better outcomes for those engaging in personalization, both early-stage and mature programs alike.
I want to uphold the need for an active role among UXers, and for architecture-as-design, in how we begin to shape the IA of AI. I’ve been struck by the need for sherpas in this noisy landscape that can guide organizations and bypass the empty hype for lasting results from holistic strategies for personalized product opportunity.
You are needed.
Ease our way further down a promising path.
Did you find this helpful? Drop a line or leave a comment and let me know. Acknowledgements are due a variety of folks, including everyone cited or mentioned above.
Appendix: Further reading
Success in a personalization initiative requires understanding the solutions landscape—and the true place of your needs…bucket.studio
The following are cited in one form or another by the post above. For those seeking a broader reading list on personalization and design, consider looking at the appendix of this post (flagged above), which I’ve updated numerous times.
Personalization is causing a seismic shift across the landscape of consumer-facing brands, and we are only starting to…www.bcg.com
One of his many observations is the growing power of ‘algorithm businesses’. By his definition Galloway sees these…www.marketingweek.com
Nothing is perfect. So even for the most successful platforms, design needs to continuously evolve. But beyond mere gut…www.fastcodesign.com
Design can benefit immensely from cognitive bias.uxdesign.cc
So much knowledge out there, so little time. Below is an attempt to outline why I’m investing most of my attention into…medium.com
Most retailers are persuaded of the value of personalization and have made at least some effort to make personalized…retail.emarketer.com
Order your favorite oven-baked goodness on your favorite devices.anyware.dominos.com
This is the prose version of a talk I gave in May 2017 at Pixel Up! in Capetown and at An Event Apart in Boston. It…bigmedium.com
Information Architecture is much more than a sitemap or wireframes. The complexity and contradiction of what people and…understandinggroup.com
Barbara Lee Fredrickson  is an American professor in the department of psychology at the University of North…www.wikiwand.com
For the last 10 years, news feeds have been the main way - the mainstream user interface - to discover interesting and…techcrunch.com
Editor's note: Jarno M. Koponen is a designer, humanist and co-founder of media discovery startup Random. His passion…techcrunch.com
Personalization algorithms influence what you've chosen yesterday, what you choose today and what you'll be choosing…techcrunch.com
Subscribe to Bucket List, the newsletter where this first appeared.