AI retention tools have advanced in leaps and bounds in recent years, enabling brokers to target client communications in a much more data-driven way. In-platform assistants and platform telemetry, combined with machine learning models, can categorize user interests, monitor engagement, predict churn, and even assess the riskiness of different trading styles. <\/p>\n\n\n\n
To be maximally effective, though, tools require proper technique. This is what differentiates their application, whether it be a surgeon’s scalpel, a stone mason\u2019s chisel, or indeed the AI client segmentation tools employed by businesses today. <\/p>\n\n\n\n
In this article, we explore what nudge theory<\/a> tells us about how to use the various user and revenue retention tools that are available to brokers. The idea is to get retention teams thinking about how to best use the data they now have at their disposal to create effective, well-timed client communications and retention initiatives. More specifically, the article offers insights into how to program different types of outreach to increase the chances they\u2019ll be received at the right moment for the client. <\/p>\n\n\n\n Nudge<\/em>, written by behavioral economist Richard Thaler<\/a> and legal scholar Cass Sunstein, explores how humans make decisions, focusing on the gulf between how we think we make them (rationally) and how we make them in practice (heuristically). Thaler went on to win the Nobel Prize in economics in 2017 for his work in the field. <\/p>\n\n\n\n Its central argument is that humans are not rational decision-makers, because they tend to employ quick-and-dirty heuristics (System 1, to borrow Kahneman’s shorthand) rather than logical reflection and analysis (System 2), which cost far more in time and energy. Taking these two modes of thinking into account, small and well-timed tweaks in how and when choices are presented can produce very different outcomes without being manipulative or removing freedom of choice. <\/p>\n\n\n\n The book focuses on health, wealth, and happiness outcomes and, as such, has greatly influenced the sphere of public policy. However, its insights have become widely used in marketing, particularly in the online world, where nudges are both easy and cheap to deliver. <\/p>\n\n\n\n Applied to client engagement in trading, nudge theory<\/a> improves user experience by timing subtle prompts against client behavior and preference data. Those prompts help users learn the product, make better choices, and address problems before they escalate. <\/p>\n\n\n\n Retention stops being a rescue operation mounted after the client has already gone and becomes a matter of serving users what they need, when they need it. Both the broker and the trader get a longer relationship out of it. <\/p>\n\n\n\n In a trading context, timing carries more weight than it does in most consumer settings. Traders keep to preferred sessions and markets, and their routines extend either side of the bell, shaped by the asset classes they favor. <\/p>\n\n\n\n A session covers the news cycle, market analysis, live observation, and the work of entering, managing, and exiting positions. Your communications should land in the gaps these routines leave open. <\/p>\n\n\n\n Real-time, in-platform communications beat emails and calls, and they work even better still when the topic matches what the client is doing at that moment. Both factors raise the likelihood of a favorable response. Account management messages, whether deposits or payment details, land best once a behavioral trigger shows that the user is viewing balances or moving funds, rather than while they are drawing trend lines. <\/p>\n\n\n\n Timed well, AI client engagement removes frictions that would otherwise surface as support tickets or drop-offs. An assistant that reaches out while a user is stuck on account verification or a failed deposit is working at the point of friction rather than after it. This smooths onboarding for the client and frees human resources, which keeps sign-ups moving at scale. Our own behavioral analytics work<\/a> identifies repeated deposit failures and difficulty placing stop-loss levels as exactly the kind of events that mark a client as receptive to help. <\/p>\n\n\n\n Nudge theory suggests that the timing of any outreach ought to take clients’ mental states, attention peaks, and decision windows into account to avoid coming across as annoying or tone-deaf. For example, it\u2019s probably better not to bombard clients with asset suggestions, educational content, or offers the second they log in to their platform on a Monday morning. <\/p>\n\n\n\n Monday morning is the time when traders want to see how the weekend\u2019s news is likely to influence the market opening. It\u2019s also the first session after the weekend, requiring some time to adjust to predominantly System 2 thinking, which is more taxing. These times should be used only to share information specifically pertinent to their activities, such as changes in margin requirements or fees that might affect their positions. <\/p>\n\n\n\n The same applies to Friday. People in general are more disengaged on Friday than earlier in the week, but traders, specifically, are also concerned with how they are about to close the week, PnL-wise. 24\/5 markets also tend to see a flurry of last-minute trades on Friday afternoons, which can bleed over into other markets as well, so it’s best not to have your messages competing with this. <\/p>\n\n\n\n The same applies on a micro level: try not to send anything at the open or close of a trading session, as these times can often be the most volatile in the session, and you\u2019re essentially asking them to split their attention from the very reason that they\u2019re logged into your platform. <\/p>\n\n\n\n Midweek (Tuesday-Thursday) is generally regarded as a useful window where you\u2019re more likely to get a client\u2019s attention than at the beginning and end of the week. This window is better served by informational\/educational content, asset suggestions, or prompts to experiment with a new feature. <\/p>\n\n\n\n Finally, consider that real-time messages have a logic of their own, so don\u2019t just copy your email blast process to this domain. For example, they should neither be sent too soon nor too late after the client logs in. Give your users a moment to get back into the swing of things before reaching out, or your message risks becoming another irritating item on their to-do list. Sending messages a minute or two after clients log in can help with the above, but it also makes messages feel more organic and relevant, since they weren\u2019t sent while the user was offline. <\/p>\n\n\n\n Defaults are important in nudge theory, as they\u2019ve been shown to be the most powerful type of nudge. Internet users are bombarded with choices and are constantly engaging with platforms trying to grab their attention in various ways, which leads to decision fatigue. The way in which you set up default settings can help to standardize certain \u201cgood choices,\u201d allowing your users to take the path of least resistance. This is because opting out of default settings takes more time and effort than just accepting them, which allows you to standardize certain best practices. <\/p>\n\n\n\n For example, automatically setting stop losses for new accounts to within a certain percentage of the current market price standardizes the practice of applying risk management to a trade in a way that\u2019s beneficial to the client, while also serving an educational function. In this way, users are nudged to engage with risk management features to change the default or remove it, and thus must learn how it works. <\/p>\n\n\n\n Furthermore, defaults serve as a form of social proof, which makes users more likely to accept them. A default setting is received as if it were placed there by someone who knows better, so it comes with a sense of authority. It\u2019s also a setting that other users presumably have by default, which makes people more likely to leave it as is. <\/p>\n\n\n\n Keeping the twin concepts of social proofs and smart defaults in mind when setting up trading assistants can be a most effective way to apply the nudge theory. For example, a beginner requesting some symbol suggestions could be responded to as follows: <\/p>\n\n\n\n \u201cMost new traders with your profile start by following the S&P 500, Nasdaq, and Gold.\u201d <\/p>\n\n\n\n Alternatively, in response to a position entry instruction, a smart assistant can also do the following: “I’ve set a stop-loss at -5% in order to protect your capital. Keep or adjust?” <\/p>\n\n\n\n Additionally, simplified default platform layouts for newcomers ease the learning curve, allowing beginners to familiarize themselves with and learn platform functions in a less overwhelming way. <\/p>\n\n\n\n The way choice consequences are presented can make a big difference in how they\u2019re understood by clients, which can impact behavior. For example, even intermediate traders will routinely miscalculate the risk they\u2019re taking depending on how an option is presented to them. <\/p>\n\n\n\n To help a trader understand the magnitude of the loss required before a stop-loss triggers, it can be presented as a percentage rather than a simple price point. However, framing it in more human terms can further improve sensitivity to risk management and help with decision fatigue. <\/p>\n\n\n\n Trading assistants are an ideal way to offer this kind of human-centric framing. For instance: \u201cWith your stop-loss set at $210, or 5% below the current market price, you stand to lose $700 before your position is stopped out. Shall I proceed?\u201d Such an approach can also be used to prevent \u201cfat finger\u201d mistakes. For example: \u201cThis position is ten times larger than your usual trade size. Are you sure you want to continue?\u201d <\/p>\n\n\n\n Helping users understand how their decisions map to outcomes in the most intuitive way possible helps reduce decision fatigue, which is even more important for brokerage clients because trading requires heavy system 2 engagement, which is taxing by its very nature. These small, thoughtful interventions can greatly improve user experience, brand loyalty, and lifetime value. <\/p>\n\n\n\n In an updated edition of the book, the authors introduced the concept of an undesirable form of nudge, which they termed \u201csludge.\u201d Also informed by behavioral economics, sludges take advantage of the phenomena described above to introduce frictions into user experience and influence a desired outcome in a manipulative way. <\/p>\n\n\n\n Nudges are considered good when they follow certain ethical principles, such as transparency and ease of opting out. They\u2019re considered bad when they include hidden costs or force users to take extra steps to opt out, hence the name sludge. <\/p>\n\n\n\n Default cookie preferences on websites can be considered a form of sludge<\/a> that exploits the default effect. They typically have some cookies already accepted by default, requiring users to press several times, sometimes even forcing them to navigate away, in order to alter and save their preferences. Adding extra items to a checkout cart by default, such as item insurance, is also considered a form of sludge. <\/p>\n\n\n\n These types of manipulative nudges are to be avoided, particularly in a trading context where you\u2019re attempting to build long-term relationships with your clients. It\u2019s generally considered best to stick to nudges that improve user experience, allow them to proceed at their own pace while maintaining engagement, and are transparent when it comes to costs and risks. <\/p>\n\n\n\n When developing AI tools for brokers<\/a>, our approach has been to focus on a combination of client-facing and broker-facing products<\/a> that work together to generate data-driven insights and deliver content informed by them. <\/p>\n\n\n\nNudge theory<\/h2>\n\n\n\n
Timing is everything <\/h2>\n\n\n\n
Be smart about your defaults <\/h2>\n\n\n\n
Understand choice mapping <\/h2>\n\n\n\n
Nudge vs sludge <\/h2>\n\n\n\n
The Devexperts approach <\/h2>\n\n\n\n