How to Discover Your App’s Retention Hook and Improve User Retention

With more than five million apps in the two-leading app stores1, app publishers are making larger investments in app development and user acquisition. Most are failing to achieve success, as demonstrated by the average app losing more than eighty percent of their users within just seven days of downloading the app.2 Will your app do better? Many try to do better by adding features to their app with an unrealistic expectation that the next feature delivered is going to engage users and improve user retention. Adding features without understanding which features users value results in more disappointed users that may never come back to the app.

Source: https://www.similarweb.com/blog/report-uninstalls-the-data-behind-deleted-apps

Ankit Jain, CEO of Quettra (acquired by SimilarWeb), provided his observation on the user retention pattern displayed in the data Quettra collected on more than four million Android app users.

“Users try out a lot of apps but decide which ones they want to ‘stop using’ within the first 3-7 days. For ‘decent’ apps, the majority of users retained for 7 days stick around much longer. The key to success is to get the users hooked during that critical first 3-7 day period.”2

Six Steps to Discovering your Retention Hook

So how do you get users hooked on our app in less than seven days? The answer is to perform a cohort analysis to discover your Retention Hook. The cohort analysis has six steps:

Step 1: Identify your cohorts. Group your users by the pattern of actions they take within a specified period of time. Invest time in this critical first step to ensure that you identify all possible cohorts. Some cohorts may not be obvious or appear to be too small to be relevant at first.

Step 2: Measure how long each cohort remains active in the app. There are several ways to measure retention. It is important to match the retention measurement methodology to the expected usage patterns necessary to achieve your app’s objective.

Step 3: Analyze the retention measurements and build a hypothesis of the Retention Hook from the cohort that remains active the longest.

Step 4: Test to validate the hypothesis on a larger group of users.

Step 5: Modify your app to guide users through the Retention Hook.

Step 6: Repeat the process to refine the Retention Hook or discover a new Retention Hook.

Discovery of a Retention Hook Fueled Facebook’s Growth Past Larger Incumbents

When Chamath Palihapitiya was leading Facebook past MySpace in user growth, he focused the team on long-term sustainable growth instead of trying to create a viral experience.3 Cohort analysis uncovered the now famous 7 friends in 10 days ‘a-ha’ moment. Tests revealed that when users achieved at least 7 friends within 10 days of signing up with Facebook, they realized the value of Facebook in their lives and were 85% more likely to return to the app. Guiding users to this ‘a-ha’ moment (retention hook) provided a clear and measurable objective to direct the actions of the user growth team.

Cohort Analysis Can Uncover Surprises that Deliver Dramatic Results

Your cohort analysis may deliver surprising answers. Calm, a simple mindfulness meditation mobile app, used a cohort analysis tool, Amplitude, to discover their Retention Hook.4 At first glance, you would think that doing a meditation exercise, the core feature of the Calm app, would be the Retention Hook. What Calm discovered was that a few users of a minor feature buried on their Settings page had a retention rate that was three times of those who didn’t set a reminder. This is why it is important to carefully consider the actions that users are taking within the app in Step 1 of your cohort analysis.

Since the group of users was so small, Calm was uncertain of the causality of this behavior, so they conducted an experiment. They prompted a group of new users that finished a meditation session to set a reminder. Forty percent of those prompted set a reminder and demonstrated a similar retention rate of three times those that did not. Calm released a new version of the app that prompted new users to set a reminder which dramatically improved their user retention.

With Only 3-7 days to Engage, You Must Know Your Retention Hook

Think of the times you have downloaded a mobile app. If the mobile app didn’t deliver an engaging experience or obvious benefit, did you keep using it? Did you keep upgrading to new releases until it did? Or, did you abandon it and delete it? Why would you expect users to behave any differently than you do?

The challenge is to suppress the urge to build new features until you understand what is causing your users to engage with your app. Cohort analysis is a tool that can help build that understanding.

 

If you would like to learn more about how discovering your retention hook would improve your user retention, you can reach Kevin Struthers at kstruthers@wcapra.com.

 

1 https://www.statista.com/statistics/276623/number-of-apps-available-in-leading-app-stores/

2 http://andrewchen.co/new-data-shows-why-losing-80-of-your-mobile-users-is-normal-and-that-the-best-apps-do-much-better/

3 https://ryangum.com/chamath-palihapitiya-how-we-put-facebook-on-the-path-to-1-billion-users/

4 https://amplitude.com/blog/2016/07/14/calm-increased-retention-amplitude/


C-store Industry Moves Beyond the Basics in Technology Investment

The following has been re-posted from Convenience Store News. The original article can be found here.
By Don Longo – 09/19/2017

NATIONAL REPORT — Almost 50 percent of retailers responding to the exclusive Convenience Store News 2017 Technology Study said they plan to add new equipment and replace technology this year, with an expected growth in capital expenditure spending of 8 percent more than in 2016.

“With per-store spending, according to the CSNews survey results, approaching $20,000 per store in 2016, one may argue this was to implement EMV inside the store,” said Ed Collupy of W. Capra Consulting Group. “However, the expected growth in 2017 technology investments tells me that people see systems as a way to bring more to their business and are less driven by compliance mandates this year.”

For an analysis of this year’s Technology Study results, CSNews turned to Collupy, a former retail technology leader and current executive consultant for the W. Capra Consulting Group, a leading retail technology, payments, security and strategic implementation firm.

Over the last several years, the convenience store industry has been transforming itself with the addition of stronger foodservice offerings and retailers have invested in preparation and display equipment, people, and remodeled stores. This year’s CSNews Technology Study indicates the realization that systems can further help grow this aspect of the business, noted Collupy.

Foodservice ordering kiosks, both inside and outdoors, show continuing and significant growth over the past few years: more than 30 percent of survey responders indicate they’ve implemented or plan to implement these customer-experience technologies both in-store (12 percent in 2016 and 9 percent in 2015) and outside (6 percent in 2016 and 4 percent in 2017).

The plumbing (i.e., network and security hosting, monitoring and management) of information technology (IT) in companies continues to shift to the “cloud.” Collupy explained that these technologies are often viewed as a necessity but not the best use of the time and resources of in-house IT professionals. Thus, management continues to outsource many of these activities. Sixty-four percent of respondents reported that their companies outsource IT functions. The study shows steady growth in this area, up from 50 percent in 2015 and 59 percent in 2016.

“In my discussions with retailers, I hear that the main driver for this is to ‘let my people focus on systems that will bring added value to the business,’” said Collupy.

One of those added-value areas appears to be social media. Social media has become an important marketing element for c-store operators, as almost three-quarters of respondents said they are using social media applications.

This year’s CSNews Technology Study also revealed an increase in new products being introduced to customers via social media. “To me, a key aspect of this technology is creating a dialog with customers,” said Collupy. “Close to 46 percent of respondents use polling features available in social media apps. In addition, those companies reporting they have their own mobile app are also using it to create a more engaging customer experience.”


Fraud Programs – Doing More Harm Than Good?

U.S. merchants had an astounding $9 Billion in payment card fraud in 2015, a number that continues to grow at an increasing rate year after year.[1] In a retail climate that that faces increasingly tighter margins, higher operating costs, and intense competition from low cost, online only retailers, fraud losses are a major pain point for organizations, especially for their bottom line.  In efforts to thwart fraudsters, especially in card not present environments (i.e. orders via mobile and web platforms), organizations have begun designing and implementing fraud monitoring programs. These programs may include a combination of both internal controls and 3rd party fraud management solutions to reduce the impact of card fraud. Unfortunately, many of these fraud solutions are more successful at turning away good customers than preventing the bad actors. Why is it that the good customers are getting turned away, but the fraudsters still manage to get through?  Our experience has shown that a disconnect between organizational goals and the process for achieving those goals may be to blame – leading to unintended outcomes and higher fraud costs.

Fraud solutions have a large thorn in their side, namely false positives. A false positive occurs when a fraud management tool or program incorrectly identifies a good customer as a fraudster and declines a transaction. The transactions that most fraud solutions target are a small subset of total chargebacks, which is an even smaller subset of all transaction volume. All of this leaves a lot of room for error when assessing transactions for fraud. Estimates range between 1-10% of chargebacks are the result of criminal fraud, the type of fraud that most fraud solutions target.[2] Therefore, only a small amount of transactions should be declined by a fraud solution. Yet in 2014, an estimated $118 billion in valid sales were incorrectly declined due to fraud solutions; 13x more than amount of payment card fraud that merchants incurred.[1] For e-Commerce/m-Commerce merchants the false positives problem is only getting worse – false positives jumped from 25% to 35% of declined transactions from 2015 to 2016.[3] It is evident that the rate at which the industry declines transactions is doing more harm than good. While false positives are an immediate concern, the long-term impact of false positives is often overlooked.

Customer experience is a key driver for sales, and poor experiences can have a devastating effect on consumer behavior. Approximately 26% of declined cardholders reduce their spending at a merchant following a decline, and 32% stop shopping at a merchant entirely (for e-Commerce transactions 66% of cardholders reduce or stopped patronage at a merchant).[1] As a customer, having a transaction declined is an embarrassing experience that causes many to leave a merchant altogether. Not only are false positives killing over 13 good sales to every fraudulent transaction, the negative customer experience has a significant impact on future sales. It is clear that false positives pose a great risk to the industry; luckily, there are some tools and tactics that merchants can adopt to help reduce legitimate fraud and minimize false positives.

Manual reviews have become an industry standard for big box retailers to lower false positives on eCommerce transactions, however, it is a costly and time-consuming solution. There are alternative approaches that are easy to implement and can be applied to merchants that can’t afford the time to perform manual reviews. The first is ensuring that any data modelling or transaction scoring models are trained on lowering false positive rates rather than lowering fraud. In addition, creating control groups to analyze the true rate of false positives will help organizations understand the harm they are causing and react appropriately. Rather than outright declining potentially fraudulent transactions, utilizing additional customer verification methods such as two-factor authentication or customer specific knowledge may reduce false positive rates. These tactics along with a shift in organizational fraud objectives will help merchants cut back fraud without reducing sales.

As an industry, the underlying goal of fraud programs needs to be re-evaluated. Simply looking at losses due to chargebacks and determining that number needs to be reduced is ineffective. In other words, fraud teams that have been established with the sole goal of reducing fraud will reduce sales even more than fraudulent transactions. When the primary goal is to reduce fraud, the solution will include a rampant rate of false positives that will have a significant effect on a merchant’s bottom line today and in the future. Progressive fraud solutions must possess a more holistic view of fraud as it relates to the business; the goal of fraud prevention being profit maximization for the organization. Combining a shift in organizational fraud goals and various methods to reduce false positives, the unintended consequences of fraud solutions, false positives, will be minimized.

If you would like to learn more about managing fraud at your organization, including how to best protect against false positives, you can reach Danny Omiliak at domiliak@wcapra.com.

 

[1] https://www.riskified.com/pdfs/Riskified-Javelin-Whitepaper-2015.pdf

[2] https://chargebacks911.com/2016-lexisnexis-study-understanding-true-cost-fraud/

[3] http://images.solutions.lexisnexis.com/Web/LexisNexis/%7Bea78e9df-056e-46ed-b04c-e8bfbc526ffd%7D_2016_True_Cost_of_Fraud_Study_052516.pdf?elqTrackId=cc6d22e0b40d4a29a2931f28fb221092&elqaid=2567&elqat=2