a16z to Entrepreneurs: Instead of trying to save low retention rates, it’s better to pivot directly.

Author: andrew chen, a16z

Compiled by: Tim, PANews

I have been keeping an eye on retention curve data for over 15 years.

I have seen thousands of retention curves, and this is one of the first metrics I ask to see when evaluating startups. I have gone through thousands of databases and analyzed retention curves broken down by different segmentation dimensions. As a product builder, I have also observed this metric from another perspective. I have run hundreds of A/B tests, drafted countless versions of user onboarding guides and notification emails, trying to change the shape of the retention curve.

A/B testing (also known as split testing or bucket testing) is a random experimental method used to compare two versions of a product (Version A and Version B). Its core purpose is to determine which version performs better in achieving predetermined goals by collecting data and analyzing user behavior.

From the results, there are some patterns here.

Just like the laws of physics, it is strange that over time, there are always certain deterministic patterns that keep emerging. Here are a few examples I want to share:

• You cannot improve a poor user retention rate. Yes, adding more notification features will not improve your retention curve. You cannot achieve good user retention through A/B testing.
• Retention rates will only decline, not rise. Interestingly, its decay rate indeed follows a predictable half-life pattern. Early retention rates can predict later retention performance.
• Revenue retention expands, while usage retention shrinks. The good news is: although users will gradually churn, the users who remain sometimes spend more!
• Retention rate is closely related to your product category. There are both innate reasons and those cultivated over time. Unfortunately, you are destined not to make a hotel booking app a product used daily.
• As user expansion and growth occur, retention rates tend to decrease. The highest quality users come from early and organic growth, while users acquired later perform the worst.
• User attrition is asymmetric; losing a user is much easier than winning them back.
• The calculation of retention rate is very difficult. Seasonal factors do exist, and newly launched test versions can interfere with the data, while system vulnerabilities also occur from time to time. Although D365 is a real indicator, one must not look at this result alone.
• Viral growth but extremely poor retention rates will inevitably lead to failure. We have repeatedly verified this conclusion across multiple platforms and categories.
• Outstanding user retention is nothing short of a miracle. When you actually witness such a miracle, you will feel incredibly shocked.

We will analyze these points one by one.

You can't save a poor user retention rate. You've seen it happen: you spend months developing a new product, and then you launch it. The initial user retention data is dismal, hitting you hard. At this point, after months of product development, how can you improve retention? Suddenly, an idea strikes you: why not add a push notification feature to remind users to come back? Or add a bunch of new features? Perhaps conduct A/B testing on the landing page to improve conversion rates?

I think we all know how the story will end. Unfortunately, when the product retention rate performs poorly, it is often extremely difficult to turn things around, and it could be said that it is almost hopeless. Of course, there may be some marginal improvements. Suppose your next-day retention rate is 40%, and the goal is to increase it to 50%, that is entirely feasible and worth the effort. But if the next-day retention rate is only 10%, it likely means that the product you have created does not meet market demand at all, and at this point, all the local optimization methods surrounding A/B testing and push notifications are not enough to change the fundamental dilemma. When months of development time and sunk costs are already a reality, it is hard not to struggle desperately. But I believe that in most cases, it is best to make a decisive choice to pivot.

This transformation aimed at enhancing user retention requires a complete redesign of the application homepage. If it was originally presented in an information flow format, perhaps it should shift to a structured step-by-step process; if the core of the product lies in sharing functionality, the focus should possibly shift towards content creation and curation. You may need to describe the product positioning in a completely different way, even aligning it with competitors. This must be a large-scale transformation across multiple dimensions, the more thorough the better, as this is the only way to potentially reverse the situation of low user retention.

The retention rate will decrease, but it will not increase. The retention curve typically presents as a very regular geometric curve pattern. For example, many curves I have observed follow this rule: regardless of what the first-day retention rate is, the seventh day will see a decrease of 50%; regardless of what the seventh-day retention rate is, the thirtieth day will see another decrease of 50%. Over time, the final retention rate may approach zero, although if lucky, it might be maintained at around 10% overall. This decay pattern is predictable.

You have never seen a curve that first rises, then falls, and then rises again; it is impossible. In other words, if the early retention rate is not outstanding, then the late retention rate is likely to be unsatisfactory as well. You must start strong in order to finish well.

There are some noteworthy exceptions in this rule that need to be pointed out specifically:

• Some products are very hardcore (e.g., online poker). The user retention rate for such products may be relatively low, but the users who stay are often extremely loyal and spend a lot, and this model has proven to be successful.
• For products with network effects (such as social networks, collaboration tools, or other products with network effects), new users may initially show high activity, but their engagement may temporarily decline afterward. However, if the product can leverage an increasing number of users to re-engage old users, there is usually a slight rebound in retention rates. This situation is extremely rare, but once achieved, it is astonishing.
• Revenue retention is expanding while user retention is shrinking. One of the best and most important characteristics of the retention curve is that it can be applied to both users and revenue. So far, we have been discussing user retention, but unfortunately, user retention tends to show a downward trend, which is not ideal. On the other hand, revenue retention is quite interesting because the users who remain tend to spend more money on your platform over time.
• This is one of the biggest advantages of B2B SaaS products. Take Slack as an example; if you observe the user group data, you will find that its retention curve shows a downward trend, just like other products. Some people accept it, while others do not buy into it. However, for those companies that invest time in deploying Slack, the product will begin to grow naturally, and the revenue you receive from these enterprises will increase over time. The revenue retention curve rises instead of falling, which is a remarkable phenomenon, but unfortunately, it does not apply to most consumer-grade products. It is this characteristic that allows B2B products to have a smoother business model compared to consumer-grade products.
• The consumer application model is closer to Amazon, where you might initially only purchase books and music, but as the product features continue to expand, you gradually start buying more and more items with it. For this reason, the lifetime value of users in the product essentially has no upper limit. We observe a similar phenomenon at Uber: although the user group may decline over time, the taxi spending that people initially used for airport transfers gradually expands to restaurant trips or commuting scenarios. Therefore, while the user retention curve shows a downward trend, the revenue retention curve continues to rise.
• Retention rate is closely related to product categories. In the past, I have discussed the innate and acquired factors of retention rates. The reality is that many products have natural usage scenarios; for instance, collaboration tools or programming software that you might use every day while working, but the upper limit of usage days is typically 5 active days out of 7 days a week. In contrast, vulnerability alert systems hope that users do not use them frequently. Consumer goods are similar; people check news, messaging, and social applications daily, but typically do not use medical reference guides very often. Some applications have low usage frequency but high retention rates, such as weather or banking applications. On the other hand, categories like games, which are addictive and frequently used, often see users drop off just a few weeks after content consumption.
• The importance of innate and acquired factors lies in the fact that they reveal the reality that many new products struggle to break through. If you are developing a social travel application, but people’s actual travel frequency is low, then creating a product centered on interaction among friends will be extremely difficult. A wiser approach is to accept its low-frequency usage attribute, enhance monetization capabilities by controlling the transaction process, or integrate high-frequency usage scenarios like restaurants and nightlife, similar to Yelp, while retaining travel features. It is indeed challenging to go against the trend, and there is very little we can do.

For this reason, if you want to create an application with a very high retention rate and usage frequency, you may need to choose fields that users already regard as core daily products for development. This means that successful applications are likely to take away usage time from other daily products, just like my frequent use of ChatGPT has significantly reduced my Google search frequency; when I started using Substack to read and write blogs, I gradually abandoned other various social news software.

When expanding the user base, the retention rate often decreases instead of increasing. Even if you are fortunate enough to create a high-retention product, people often instinctively extrapolate and directly apply the behavior patterns, monetization capabilities, and usage habits of existing users to a broader market, believing that as long as they multiply some good small data with core big data, they will naturally achieve extremely impressive macro results. But the reality is often: as the user base grows, problems begin to emerge. For example, when you start to expand into Android users and international markets, acquiring more customers through paid marketing and other channels, you will quickly find that there is a decline in all key metrics.

The reason is that high-quality users often appear earlier. Those user groups with the greatest monetization potential, strongest willingness, highest degree of digitization, and most active online behavior usually start using the product early through recommendations from friends. As new users are acquired later from other channels, the product may not align as well with their needs. For example, if you develop an iPhone application for university students in Western countries, when expanding to Android users in emerging markets, the functionality may not be fully compatible, leading to a natural decline in various metrics. Although continuous optimization and improvement can be made later, I can assure you that the results will never be comparable to those of the early user group.

So, the question arises: as the number of users grows, the quality of users gradually declines, do they still hold value? Can the product continue to be profitable? More importantly, can the core group of high-value early adopters be retained?

No wonder these early users are often referred to as the "golden group".

User churn is asymmetric. It is extremely easy to lose users; in fact, most products lose 90% or more of their users within the first 30 days. At the same time, it is very difficult to win back users who have already churned. This asymmetry between acquisition and churn is the core characteristic of user attrition. The reality is often so bad that rather than trying to regain old users, it is easier to acquire new users directly.

For this reason, trying to awaken dormant users through lifecycle marketing by sending discounts or promotions is often costly and yields little effect. A more effective approach is to allow existing active users to awaken dormant users through the natural use cases of the product. For example, when a professional fails to continue using a new project management tool after trying it out, sending bombarding reminder emails to their inbox is unlikely to win them back. A more effective approach is to have their colleagues invite that user to re-engage with the tool for a new project, which is the effective way to do it. However, this strategy is extremely challenging and complex to implement, and typically only products with network effects (i.e., shared and collaborative features) can adopt it.

Retention rate is very tricky and difficult to measure. When people talk about retention rates, they often tend to measure the first day, the first week, and the first month, but rarely discuss what happens two years later. This is because, during product development, teams need a sufficiently short time span and easily measurable metrics to make decisions based on them. Therefore, although annual user churn rate or long-term monetization ability is extremely important, people often do not measure these and instead focus on the easily measurable metrics at hand. However, this approach has many problems.

Unfortunately, many product categories are strongly affected by seasonal fluctuations. E-commerce, travel, health services, and online dating are typical examples. Even the way businesses use commercial software can have cyclical variations. Seasonal factors can interfere with judgment; you may find that monthly or quarterly data declines, but is this because the newly launched features are unpopular? Or is it that user behavior patterns in this quarter are inherently different? It is indeed difficult to make effective assessments when retention data is severely lagging.

Similarly, whether it is program vulnerabilities, newly conducted tests, or new marketing activities, these factors can disrupt data. Ultimately, you will find yourself constantly referring to reports that show fluctuations in retention curves, but each data point comes with additional notes, as the team needs to verify whether the newly launched Android version has led to irrelevant comparison situations.

The explosive growth in users combined with a dismal retention rate is destined to fail. Many new product developers often focus excessively on new user registrations while completely neglecting user retention. After all, if all you want to see is a continuously rising graph, why not simply increase the traffic at the top of the funnel to showcase rapid growth? After raising a large amount of venture capital this way, it wouldn’t be too late to slowly address the user retention issues.

This phenomenon is not uncommon in the current industry: a creator promotes their app to millions of followers, or a single video post leads to a surge in revenue, and the product achieves a spike in users through TikTok. Although the actual usage rate and user churn situation are not ideal, this phenomenon continues to occur.

The technology industry has undergone countless experiments like this. The conclusion is always the same: products that go viral but have poor user retention will eventually disappear, as the retention issue is hard to solve. When the novelty wears off, user acquisition will slow down, and ultimately, you will face a bleak situation of both user acquisition and user retention. The higher you climb, the harder you fall.

We have witnessed this phenomenon in many scenarios. In the early stages of social networks, many products achieved growth by crazily sending spam emails by obtaining users' email addresses and contact lists, but ultimately led users to low-quality products. Sometimes, as long as users can subscribe to certain low-quality ringtone annual services, companies can try to monetize from it and make a profit. However, it wasn't until the emergence of Facebook, with innovations in user experience such as information feeds and real-name systems, that a product with high viral spread characteristics and strong user stickiness was finally created. The same situation occurs in the mobile application field, where we sometimes see applications that suddenly become popular by forcibly inviting users via SMS, but if the product lacks stickiness, the entire model will quickly collapse.

High retention rates are simply magical. You might feel a bit frustrated after reading this article; I know that launching a project can sometimes be really tough. But when a product truly works, that feeling is unparalleled. When you witness a product achieving a 50% 30-day retention rate (I've seen this happen once every few years), the shock is hard to describe. I gradually realized that these fleetingly successful products are not the result of the creators having a systematic A/B testing methodology, nor are they achieved through high-speed iterative processes. The real key lies in that touch of magical inspiration. This magic comes from breakthrough insights into market or customer needs, which may seem obvious in hindsight, but allows the product to achieve exceptionally high retention rates by being the first to realize this understanding. This is true when we evaluate video conferencing software, self-destructing photo features, or magical artificial intelligence that can respond to any topic; this magic cannot be obtained solely through iteration and metric-driven testing.

The real question

You may still have a big question after reading all of the above: Wait, so how exactly can we achieve a high retention rate? (If I could answer this question with certainty, my job as a startup investor would be so much easier, wouldn't it?)

But let's do our best. In my above points, there are actually some clues buried: ideas are really important.

If you want a product with a high retention rate, you need to choose a category that inherently has a high retention rate.

You need to choose a product category that you are already using on a daily basis.

You will build a product that directly competes with it.

If you win, then you will stop using that product and switch to using your own product.

This is a high demand, but I think thinking this through is a good start.

Of course, if the product you are creating directly competes with existing products, you might wonder: "It’s really difficult to get users to switch camps." And that is indeed the case. Therefore, at this point, you need to decide to take on enough market risk, but it must be a moderate risk, by launching a novel and unique product to redefine the core interaction model. However, the innovation mentioned here is more likely to refer to a 20% improvement optimization rather than an 80% disruptive innovation. Ideally, you should be able to help users quickly and intuitively understand this innovation within the first minute of their use.

At this moment, we cannot avoid one of the most frequently asked and hardest to answer questions by investors: "Why is it feasible now?" Because your answer must point out that there has been a new trend in the industry, such as the emergence of general technologies like large language models, or social changes like the oversaturation of social media, making your innovative idea timely.

This can allow you to quickly seize the existing market and is more likely to achieve excellent user retention rates in the early stages. Timing is crucial. If you miss the right timing, enter a low-attention area, and the product differentiation is not prominent enough, you will find that you have simply transformed the user retention problem into a user acquisition challenge. The difficulty of developing a new type of web browser lies in the fact that once successful, user stickiness will be extremely high. However, people are very satisfied with the existing browsers, and getting users to try a new product requires costly and complex efforts.

This is why I don't blame those who propose ideas like "Cursor for a certain field" or "Figma for a certain industry," similar to the past concepts like "Uber for a specific vertical." They are trying to leverage existing markets and behavioral patterns to avoid significant market risks.

If we can accurately grasp differentiated advantages, seize market opportunities, meet the needs of a vast number of users, and pinpoint the core product positioning, then this model can indeed succeed.

How to open new markets?

The opposing view is that new markets are often more exciting than existing ones. Shouldn't the technology industry be about building entirely new things rather than innovating 20% on old ones? Of course, this statement is not wrong, but I believe that such products account for only a very tiny fraction.

My rebuttal to this is: in fact, most products inherit some kind of "old thing," even if those predecessor products are quickly forgotten.

Before Instagram, there was Hipstamatic, which was the top paid photography app in the App Store early on, confirming the huge market potential of filter functions. Just as Google was not the first search engine (it was actually the tenth entrant, following platforms like Lycos, Excite, and Infoseek), these cases demonstrate not only the strong demand from users for search functions but also reveal the commercialization dilemmas faced by early search engines. Tesla is not the pioneer of electric vehicles, nor was the iPhone the first smartphone. History repeatedly shows that the market landscape is often determined by the tenth-generation innovators. This phenomenon is known as "latecomer advantage," and I find this perspective very enlightening.

However, true innovation does sometimes happen. The birth of Uber transformed the existing offline taxi-hailing behavior into an online application, rather than being based on an already successful ride-hailing app (at that time, Lyft was just a peculiar bus booking service). Looking at ChatGPT, OpenAI took five years from concept development to the true rise of its third version, during which there was no existing blueprint available for reference. This kind of innovative journey is remarkable and serves as a driving force for the thriving development of the technology industry, as it creates entirely new product categories at the cost of taking real risks.