seen from United States
seen from Canada
seen from United States
seen from United States
seen from United States
seen from United States
seen from Malaysia
seen from United States
seen from United States
seen from United States
seen from Germany
seen from Japan
seen from United States
seen from United States
seen from United States
seen from Malaysia
seen from United States
seen from Germany
seen from China
seen from United States
Groupe cosmique de Galois
Groupe cosmique de Galois
https://ncatlab.org/nlab/show/cosmic+Galois+group
http://www.alainconnes.org/docs/slidesgaloisihp1.pdf
https://m.youtube.com/watch?v=lXKHYW7idbg
https://m.youtube.com/watch?v=PlJIECqRZRA
Trois vidéos suivantes de Francis Brown dans la colonne de droite
View On WordPress
Groupe cosmique de Galois
Groupe cosmique de Galois
https://ncatlab.org/nlab/show/cosmic+Galois+group
http://www.alainconnes.org/docs/slidesgaloisihp1.pdf
https://m.youtube.com/watch?v=lXKHYW7idbg
https://m.youtube.com/watch?v=PlJIECqRZRA
Trois vidéos suivantes de Francis Brown dans la colonne de droite
View On WordPress
Groupe cosmique de Galois
Groupe cosmique de Galois
https://ncatlab.org/nlab/show/cosmic+Galois+group http://www.alainconnes.org/docs/slidesgaloisihp1.pdf https://m.youtube.com/watch?v=lXKHYW7idbg https://m.youtube.com/watch?v=PlJIECqRZRA Trois vidéos suivantes de Francis Brown dans la colonne de droite
View On WordPress
3 Ways Well-Known Brands Could Use A Chatbot
Interest in chat bots is rising as brands discover new ways to utilize the technology to connect with their consumers. That said, thus far, the use cases for chat bots have been limited to basic customer support functionality or humorous/cringeworthily conversations with a robotic human “personality.” For brands to create bots that are less novelty and more substance, they’ll need to figure out how the bot is actually bringing value to the user experience. See below for three different ideas for how brands can use a bot to improve user experience.
1. Suggestion Bot helps A&E Improve Content Search-ability
A&E has a great mobile app that allows users watch the cable network’s content on a mobile device. But like all products with tons of content to choose from, users can spend more time searching for content than actually consuming it.
Currently, the app’s search functionality helps users looking for specific shows, but this tool won’t help users who’re browsing. To help users browse content efficiently, A&E could introduce a helpful show suggestion bot. In this example conversation, the bot asks the user if they’re interested in the “most recent,” “most popular,” or “classic favorites.” If the user selects “most popular,” and the bot would make some recommendations.
There’s nothing show-stopping about this chat bot, but it does provide a simple and easy way for the user to navigate the app. And as a benefit to the brand, the bot conversation can be managed on a web dashboard, that way, every time new content is released the bot can update without altering the app’s code.
2. Web Bot to accompany Pepsi’s “Pepsi Pass” App
Pepsi has an app called Pepsi Pass that allows users to collect points by taking selfies and pictures of Pepsi bottles (fun!). Those points can then be exchanged for prizes like headphones or concert tickets. While the app feels like a complete experience, the accompanying website appears to have been an afterthought. Case in point: the page directs users to “Hang out. Enjoy Pepsi. Get Points. Repeat.” but they forgot to mention why -- (it’s about the prizes right?!?).
To create a more engaging experience for this landing page Pepsi could introduce a bot to hype up this month’s prizes. The bot could ask users what type of events they’re interested in, whether it be music, sports, or live events, then ask users where they are located, and finally show them all of the fun ways to redeem Pepsi Pass points in their area. Sign me up!
3. Pre-flight checkin bot for Cheap O Air
The travel industry has been quick to adopt chat bots, some with greater success than others. You’re likely to learn more from your daily horoscope than Julie the Amtrak bot, but the KLM bot is a pretty slick way to get your boarding pass. I think Cheap O Air could be the next to add a helpful bot to the mix.
Imagine there’s a change to your flight 6 hours before departure. Cheap O Air sends you a push notification to let you know there’s been a change. After opening the push notification, the pre-flight bot could offer to complete some helpful tasks such sending as an email to your hotel or car rental agency letting them know your flight has been delayed. No car or hotel, the bot can direct the user to airline customer service, cancelation policies, or weather reports.
3 strategies for using user geography in marketing campaigns
If you’re a product manager or work in app marketing you’ve probably spent some time looking at the geographic distribution of your users with a mapping product like Google Analytics or ROKO Mobi. While these map tools easily satisfy the curiosities of geographically inclined people like me, they can be written off by others as simple visuals with little actionable information.
On the contrary, keeping tabs on the geographic distribution of your users can be extremely beneficial to optimizing marketing spend. Here are three ways to use geography data to lower customer acquisition costs and increase the efficiency of marketing campaigns.
1. Spend your marketing dollars in the right places
The old rule of thumb is that consumers need to see or hear about a product seven times before really remembering the pitch. With that in mind, it makes sense to advertise your application in a market that already has some exposure to your product. For example, let’s say you have a large cluster of users in Birmingham Alabama but only a couple of users in Memphis. If you decide to buy advertising in Birmingham, the likelihood that a consumer viewing your ad has seen or heard of your product via word-of-mouth is dramatically higher than a similar audience in Memphis. Seeing your ad could just be the seventh interaction a prospective user needs to finally make the download. Leveraging your product’s existing market penetration ensures that you get the most from your marketing dollars and ultimately reduce user cost per acquisition (CPA).
2. Track the success of your regional marketing campaigns
Depending on the channels on which you’re spending your advertising money, you’ve likely picked certain regions to run your ads and you’re probably receiving information about the number of impressions or downloads you’re getting in those areas. That said, the data from any one of those sources is never complete. Someone could see your Facebook ad while at work and download the app at the bus stop on the way home. Someone else could download your app from your Instagram ad and then tell three friends who download it too. In both of those highly plausible scenarios, the tracking tools provided by ad networks fall short in providing you the full picture of how effective your marketing has been. While Facebook and Instagram will provide you the number of seconds a user “engages” with your ad, or how many download it from the ad, but they have no way of showing you how your ad catalyzed user adoption in a region. The organic growth that comes from a good ad campaign should count towards your CPA as well.
3. Determine the locality of your best users
OK, so you’ve launched your ad campaign in a region that makes sense and you’ve assessed its effectiveness, but we can’t stop there! Since not all new users are created equal, it’s important to know where your best users are located. The lifetime value of your users can be calculated in several ways. If it applies to you, the amount of money spent in your app is obviously a key indicator of value. But there are many other reasons one user could be more valuable than another. Let’s say that over the past month, you’ve acquired 1,000 downloads in San Francisco and 1,000 downloads in New York. For your app, the most valuable users share at least three pieces of content. Of the San Francisco downloads, 350 have shared three pieces of content, but only 150 New Yorkers did. In this scenario, those who downloaded in San Francisco are 20% more likely to be the highly valuable user you’re seeking. In this scenario it could benefit you to spend more marketing dollars in the bay area market to acquire more of those high value users.
As a product manager at ROKO Labs, I’m constantly working to build data visualization tools that help entrepreneurs better understand their users and grow their app business. If you’re interested in maps, app marketing, or have questions, send me a note or comment here!
Ace your next investor pitch with fool-proof user growth accounting
Whether you’re raising your seed round or series C, you can bet that investors will be asking you about user growth. Strong user growth numbers prove your product has achieved some degree of product-market fit, and if growth is sustained over a long enough period of time, investors will have a hard time disputing that you’ve got a compelling product on your hands. While showing “user growth” can be as easy as furnishing an upward trending bar chart, VCs will expect the complete picture so let’s make sure you’re prepared.
At the most basic level, if your app is acquiring new users every month, you can make the case that you’re seeing “user growth.” That said, it’s best practice to look only at active users because those who are engaging with your app are infinitely more valuable than those who aren’t. The graph below shows what a monthly active user (MAU) graph might look like for a consumer app. You can see that the number of MAUs is increasing by about 3% month over month. This app could be a promising investment based on the data shown, but prudent investors will ask that the data be broken down farther.
Johnathan Hsu is the head of data science at the Venture Capital firm Social Capital, which makes him responsible for looking critically at the user data of potential investment opportunities. In his article “Diligence at Social Capital,” Hsu explains that the simple MAU graph shown above would not do enough to convey the full story to investors at Social Capital. According to Hsu, in-depth growth accounting requires breaking MAU data down further.
At ROKO, we’ve adopted a similar outlook at visualizing app growth. The example of our “Net Growth” graph below demonstrates what accounted for positive or negative changes to the monthly active user count. To visualize this, we’ve made the following enhancements.
Eliminate retained users to isolate data that contributed to a net gain or loss MoM
Divide users contributing to month over month growth into two groups:
New Users (blue): those new to the app during that month
Resurrected (green): previously inactive users who have returned to the app in the current month
Add a third group
Churned users (red): Previously active users who have become inactive this month
Why are churned and resurrected users important? Let’s tackle churned users first. As Hsu explains, “most consumer applications don’t have a very strong mechanism to bring users back month after month…the dynamic each month in a consumer app is typically add a bunch of users and simultaneously lose a bunch of users.”
The graph above illustrates that; you see a high number of new users each month with a correspondingly high number of churned users. This app is barely growing because it is struggling to retain new users. What would be more appealing for an investor is an app that gains a smaller number of new users each month, but sees a lower churn rate. That data would prove that new users are finding sufficient value in the app to return month over month.
The second data point introduced in our more detailed growth accounting graph is the count of users resurrected month to month. Seeing resurrected users can indicate that tactics intended to bring back lapsed users are working. If you see a spike in resurrected users after you deployed new features, created new promo codes, or launched an ad campaign, you have evidence that your resurrection campaigns are working. Consider a scenario where 80% of your app’s users are no longer regularly active on the app, if even a small portion of those can be convinced to return through a resurrection campaign, you’ll dramatically increase MAU counts and prove that your lapsed users are not completely out of the picture yet.
By integrating with the ROKO Mobi SDK, you’ll have access to user reporting tools like the net growth graph pictured above. As a product manager at ROKO Mobi, I’m constantly working to build data visualization tools that help entrepreneurs better understand their users and grow their app business. If you’re interested in using our user reporting tools to better prepare for conversations with investors or have any questions about our products or how you can use them, please send me an email!
Understanding Your Users Through Segmentation
Applications are designed for people. So whether you’re a product manager designing new app features, or a marketer working to engage with customers, having a deep understanding of the people using your application is a fundamental competency. After all, the more you know about the people using your app, the easier it becomes to provide these users with useful services and desirable products. But as your user base grows, it becomes increasingly difficult to make sense of the behavior and motivations of this diverse and faceless population.
To prevent yourself from losing touch with a growing user base, it’s key to remain informed by conducting comprehensive user research on a regular basis. Unsurprisingly, this is difficult for many startups. Even the most prudent product and marketing teams can become overwhelmed with the time and man-power needed to perform sophisticated user research on a regular basis. Fortunately, user segmentation is an effective tool that can make user research and targeted marketing as efficient as it is important. In this post, I’ll explain how you can utilize segmentation to glean new perspective on your users and how to use this information to increase user engagement with your product.
Why use user segments over other research methods?
Nearly every startup how-to guide ever published will expound the necessity to keep talking to your users. Startups generally first approach this task by interviewing their users directly or through surveys. While hearing directly from your users is extremely important, interviews and surveys are time consuming and prone to bias. If your application has 100,000 monthly active users, you’d have to survey at least a few thousand of them to get a representative sample. You’d also need to conduct these surveys or interviews frequently to remain informed of any changes in user sentiment as you release new features and acquire new users. Committing resources to that kind of research can be a challenge for even the most well-funded startups.
User segmentation can provide similar insight with a fraction of the resources needed for interviews and surveys. What makes this possible is the fact that segments are created using data that you should already be collecting on your users, this data can then be used to spot trends in a user characteristics and behavior. Furthermore, because most segments are created without direct interaction with the user, the risk of bias is significantly reduced.
If you’re interested in segmenting your users, but don’t yet have a way to access the data with which to segment them, you’re in luck, because it’s not hard or expensive. ROKO Labs, among others, offer free plug-ins or SDKs, that when integrated with your application’s code, automatically begin collecting relevant user data and then visualizes trends on an easy to use dashboard.
Analytics dashboards, like the ROKO Mobi dashboard shown below, allow product managers and marketers to divine the most important aspects of an app’s health in an instant. High level metrics like new user acquisition, user churn, and time spent in app provide a decent snapshot of an app’s overall health. It’s user segments, however, that give teams the ability to analyze what user properties and behaviors are driving those dashboard statistics. And it’s understanding that “why” that will help you better address user needs and ultimately grow your user base.
Segmentation in a nutshell
A segment is a subset of your total user population that share similar characteristics. For example, a segment can be created for all users running your application on iPhone and another for all users running it on Android. This is a property-based segment, which are not tied to user behaviors. Other properties could include user time zone, current app version, or cellular carrier.
Property based segments can be interesting if you notice large engagement disparities between users with different properties. Are Android users half as likely to convert as iPhone users? If so, you might compare the quality of your Android application to the iOS one. Are users in New York spending more time in the application than users in California? Perhaps your content is better suited to the East Coast or there’s a competing app that’s more popular on the West Coast. While we cannot say for sure why engagement is higher in one segment over another, segmentation allows the researcher to identify differences in user properties that could result in diverging user behavior.
If property based segments group users by characteristics that are unlikely to change day-to-day, behavior based segments lump users together by how they’re engaging with your app. Each time a user interacts with an application, that behavior is logged as an individual event. When a user downloads your app, that’s an event. When they create an account, that’s another event. Each time they open the app, buy something, open a different screen, close the app; those are all events. Because each interaction is its own event, an app administrator has a variety of options to create segments based on a specific data point or set of data points.
Behavior based segments are best at providing insights into how engaging your application is, and whether or not that engagement is improving over time. For example, let’s say that all users who’ve logged an event on your application in the last week are considered “active.” Now let’s say that users who’ve been absent for a week but active within the last month are considered “at-risk.” By creating segments for both of those user groups, it becomes easy to see how many users are actually engaged with the application at a given time by comparing what percentage of all users fall into either group. One week later, if a larger portion of your users are included in the “at-risk” segment, there is reason to consider taking steps to retain those remaining active users before more fall into the “at risk” segment and potentially stop using the application altogether.
User Segmentation on ROKO Mobi
If you’re new to user segmentation, the ROKO Mobi platform comes with several important segments already created. Once an application is integrated with ROKO Mobi, users will begin to sort themselves into one of five behavioral segments like the “active” and “at-risk” users mentioned before. The platform also has a segment for “lapsed users,” or the those who are least likely to return, but could provide valuable feedback. A segment for “New Users” and “One-and-Done” users rounds out the behavioral segments that Mobi provides right off the bat.
The segments mentioned above have been given set parameters determined by the ROKO Labs team, however, we know that not all applications expect “active users” to return each week. That’s why ROKO Mobi provides the ability to customize user segments to best fit the specific characteristics of your application. If you run a gaming application whose users tend to visit daily or near-daily, 5 days with no activity could be cause for concern. Under these circumstances, we would suggest changing the “active user” segment to include all users active within the last 3 days and adjust the “at-risk” segment to include those inactive for 3 days or more. Conversely, a travel app might expect users to visit monthly, in which case it would make sense to expand the “active user” segment to include all users active within 30 days.
Changing segment parameters on the Mobi platform is easy, all one has to do is edit the clause that creates the segment. As you can see in the first filter below, the default formula for the active user segment includes users who have been active in the last 7 days. To change the parameters of the segment to include users who have been active within the last two weeks, just change the number of days from 7 to 14. The “Users Preview” section below shows you how many users are included in the segment based on the parameters you’ve set above.
If you’re new to user segmentation and have questions send me an email! If you’re already a user segmentation expert and have strategies you’d like to share, we would love to hear from you and find out what works for you!
Follow us for more amazing history - History You Didn’t Know
