6 design highlights to make fission activities successful!

Recently, a friend who works in operations complained to me that he wanted to do a fission activity , but when he put forward requirements for the product, he was dissed for a long time, saying that the design of many links was illogical and that R&D would definitely reject this requirement. In order to pass the review, she revised the plan 4 or 5 times according to the logic of the product, and finally it passed.

She said helplessly: Every time I want to add a new feature, I doubt my life. As an operator with no technical knowledge, it is really painful!

Indeed, usually when we want to organize an event, the process is basically that operations put forward requirements to products, and products put forward requirements to technology. After a series of communications, one event after another is launched. However, demands raised from an operational perspective often overlook some technical issues, and conflicts in the operations-product-technology food chain mostly arise from this.

Today, we will take the “Invite Friends to Receive Cash” recently launched by Ele.me as an example to see what technical details should be considered for a fission activity, so that when everyone is doing similar activities, they can reduce some communication costs and complete tasks more efficiently!

1. Activity Participation Flowchart

1. Click or scan the QR code to enter the page flow chart:

Highlights:

1) Detect the browser identifier and request the user to authorize WeChat or directly enter the login state. This function can reduce the user's operation cost, especially when everyone is using QQ/WeChat. If a link to Ele.me is shared, a complicated registration and login operation is required every time you log in, which will inevitably reduce the number of users who actually register.

2) Graphic verification code for abnormal account. I understand this verification code as a way to prevent fraud. Automatic login and automatic operation can be achieved through scripts or machines. However, adding the verification code will not have much impact on users, but it will reduce the efficiency and increase the difficulty of machine operation (some require recognition of graphic verification codes);

3) Determine whether it is the user's own invitation link. The page generated by this judgment can identify the user's situation on the one hand, and on the other hand, even if the user generates the link himself, he can still see the corresponding content on the page (the page prompts the invited activity).

2. Participate in activities and invite others after entering the page

Highlights:

Detecting the browser identifier is similar to the previous one, but the detection here is mainly for the convenience of identifying user sharing. If it is in WeChat or QQ, the corresponding operation can be presented. For example, WeChat will prompt to share with friends or share to Moments in the upper right corner, while QQ will long press to share to friends and space, etc.

2. Event Design Highlights

Note: The following analysis is based on my personal knowledge framework. Please forgive me if there are any omissions.

Took a screenshot after being invited (to be honest, I was a little tempted when I saw all these coupons, it felt like I had gotten the 76 yuan for free).

1. Both parties can get rewards:

User A is an old user, and user B is a new user. User A invites user B and shares the link with B on WeChat. User B registers and completes the first order with a payment of more than 7 yuan within 7 days of registration, and completes the second order with a payment of more than 10 yuan within 14 days. In this case, user A can get a cash reward of up to 15 yuan, and user B can get a 3 yuan store discount red envelope and an Alipay red envelope (new user registration required);

Implementation method:

The refer_id in the link, which is the ID of the inviting user, is used to determine who invited the user. Then the user behavior is monitored (there may be other ways to promote orders and retention in the App, such as sign-in, red envelopes, etc.). The user completes the order, whether it is within the specified time and whether the specified amount is met.

At the same time, according to the user's progress, "pending order", "on the way", etc. will be displayed in the background of user A, the inviter.

2. New user invitation effectiveness rules:

User A can only get the reward if user B completes the order within the specified time. There are two orders here, with a time limit of 7 or 14 days respectively, which means that the user paid for at least 2 completed orders for 15 yuan, which may also include the user's 7-day and 14-day retention rates;

The coupon for new customers is 78 yuan. For example, if the maximum discount is 15+10, you need to buy 69 products, and you need 3-6 delivery fees, which means you actually pay about 44. If it is 15+3, you need to buy 25 yuan, and the delivery fee is 3-6, so you need to pay about 10 (not considering the full discount for the time being). It is relatively easy for new customers in general stores to find a coupon of more than 7.

Implementation method:

Machine reminder + manual reminder.

Manual reminders refer to reminders from the inviter. Generally speaking, in order to avoid losing users (for example, forgetting to place an order after registration), the inviter will ask the user to place an order as soon as possible after registration.

Machine payment reminder is its own payment reminder system. If the user's payment reminder is not effective or the user does not place an order immediately due to the lack of payment reminder (for example, not on the same day), the system will remind the user through App push + SMS, pop-up reminders within the App, etc.

3. Rules for generating and issuing coupons for old users:

The coupon is valid for 7 days, so if you want to save the 3 yuan, you need to complete an order within 7 days.

Here I tried to find a nearby Burger King store. One order required at least 20 product fees + 6 delivery fees - 3 coupons = 23 yuan (so it was definitely enough to pay 7 yuan. Even after deducting the 15 yuan new member bonus, there was still 8 yuan, just over 7 yuan. For the second order, users can choose the store by themselves. Even if there is a discount, it is generally difficult to be less than 10 yuan with the discount);

And this store is 2.6 kilometers away from me, which is considered a medium-distance delivery; so there are two things involved here: which store's coupons should be given to users based on location, how much the coupons should be worth, and calculating how much the user needs to pay should be the easiest way to guide them to place an order.

Implementation method:

The flowchart shows that the user is detected, the location is determined, and then the coupons are issued according to the coupon calculation rules;

Here you may need to count user data behaviors, use algorithms to calculate and predict the stores and amounts that users are most likely to order, and then generate the corresponding ones; (This is explained in the three-lesson data analyst P1 course, such as employee turnover rate analysis and bicycle user score prediction)

For example, here I calculated that the most expensive first order is 23 yuan, and the cheapest is 8 yuan, which means it must be greater than the actual payment of 7 yuan, so you just need to urge payment.

4. Award-winning push notifications on the event page:

The push notifications here are all included in general event pages, mainly to tell users about existing winners and their winnings, to promote user participation in the event. For example, the push notification will prompt that the current xx has received xxx amount of cash;

Implementation method:

You can randomly generate usernames (or find them from other places) + randomly set the amount range (set the probability of different amounts appearing, for example, set the probability of large amounts high), and then present them on the activity page.

5. Recommendations for address book permissions:

Ele.me may have thought that users are not sure who around them has not registered with Ele.me. After all, Ele.me’s current coverage is not small, so it has come up with a solution to help users: the App directly reads the user’s address book. For those who have not registered with Ele.me, a list will be provided, and users can make targeted invitations based on the list (I have made a special comparison here for clarity).

Implementation method:

Read the address book permission, upload the address book through the API for matching, return yes if it is a new user (for example, it returns True if it is a new user), and then add it to the invite queue of "My Friends";

iOS is an invitation to send WeChat messages, I haven’t tried Android yet, but if the App is a bit more rogue, you can directly click on the invitation to jump to sending a text message (Android can directly set the content and send it directly).

6. Minimum withdrawal amount limit:

The minimum withdrawal limit for Ele.me is 10 yuan, the maximum withdrawal limit is 10 yuan, and withdrawal can only be made once a day. What does this mean?

Here I simply invited a few users, the first two are old users, and the last one is a new user. I estimated that the highest is 15 yuan. It is very likely to be in the form of "5+10", with 5 yuan for the first order and 10 yuan for the second order. That is to say, if a user wants to withdraw cash, he must complete at least the first order + second order of 1 new user, or the first order of 4 new users, or the first order + second order of 1 new user + the first order of another new user;

At the same time, if the user invites n users on the first day, and the rewards brought by n users are greater than 50, the user will continue to open the App to withdraw cash on the second day;

The reward for the second order will encourage the inviter to complete the second transaction with the new user as soon as possible. At the same time, the remaining withdrawable amount will encourage the inviter to continue to invite users (although the sunk cost is not a cost, it will still make the user continue to consider the impact of consuming the silent cost).

Implementation method:

This is actually a task wall. The product developer wants to retain users and make orders, so he gives the inviter two indicators.

It is like a master-apprentice system, where the master invites the apprentice, and the apprentice needs to complete Task 1 or Task 1 + Task 2 (for example, real name + order) and then the master gets the reward. It is also like a task wall, which is used to meet the retention needs of the product side. After the user completes the download and activation on the first day, the user will be given a second reward based on 20% of the current reward to encourage the user to continue opening the App the next day.

Appendix 1: About the strategy of making money

I don’t really think this is a highlight, but I feel that it must be explained. It is like the necessary operating procedure instructions for inviting people, and it should have been there when designing the activity.

The reason I bring this up is because some activities were not designed with this in mind.

During normal design activities, it is very easy for products or operations to find that something is simple, but users cannot perform it. Therefore, it is very necessary to have a graphic or video guide to tell users the process and simplify things. Moreover, in order to avoid accidents as much as possible, general activity rules will be accompanied by a bunch of explanatory rules and explanation elements, but users will not read them one by one.

Implementation method:

Provide a graphic description of the operation process with simple steps. The core steps must be described in detail, such as requiring the invitee to register via a QR code or link.

At the same time, if resources are in place, each step of the operation can be explained in detail through screen recording instructions, so as to minimize the cost of the inviter's participation in the event and the communication cost between the invitee and the inviter during the operation.

Appendix 2: About the invitation title and icon

The examples presented above are Ele.me’s WeChat Moments invitations and friend invitations, both with icons and promotional titles. This is also a place that is easily overlooked like Appendix 1. Good copywriting titles and icons can increase the conversion rate of users clicking on links. Only with the first step of clicks can there be subsequent conversions, especially the action of sharing in WeChat Moments.

Implementation method:

Set up multiple sets of forwarding copy titles and icons, communicate with technical staff to conduct A/B testing, and launch the best solution after obtaining it; you can also continue to optimize while doing it online to increase the click-through rate of the page.

3. Some extended thoughts on this activity

Finally, let’s talk about anti-cheating solutions in invitation reward activities.

1. Ele.me's request and response parameters

Here I checked some parameters on the Ele.me page. The screenshot below is a screenshot of the request parameters:

The above picture is the parameter diagram of the request link. For the convenience of presentation, I changed # to &; the from parameter is the source, which should be WeChat+QQ+link, etc. According to the link parameters, it is "single message". This request link is copied from WeChat (I clicked on the invitation link shared by others, and the copied one was changed after clicking);

The isappinstalled parameter indicates whether the App is installed. Since the App needs to be opened after registration, the link directly determines whether the App is installed. However, the actual judgment is not accurate. I installed the App but still opened the App Store.

The refer_id parameter is the ID of the inviter, which is used to determine who invited the invitee;

The refer_channel_code parameter is the invitation channel code, and the refer_channel_type parameter is the invitation channel type. I don’t know the details here, but I guess the channel code is WeChat/QQ/link, and the channel type is, for example, personal invitation/friend circle invitation on WeChat;

The above are the parameters of the invitation link, and the following are the parameters of the page request.

I took a screenshot of the source and whether it is opened directly on the page parameters. The request link carries the user's current access status, inviter ID, etc. (the screenshot shows the return parameters), and returns whether the user is currently opening the App directly, etc.;

2. Thoughts on the Ele.me solution

Since most of these solutions are operated based on the web after sharing (register on the web, place orders on the App), it is not easy to count the user's device, especially the unique identifier, so judging by the user's backend behavior becomes a better solution.

For example, Ele.me proposed: A new Ele.me user refers to a user whose device number, mobile phone number, Ele.me account (including third-party accounts) and payment account have never placed an order on any Ele.me food delivery platform.

Finally, if we want to do a very strict anti-cheating, we must first explain it in the rules, and secondly, the technology can be divided into the following four aspects:

1. Detect user IP, device information/browser ID; if a device registers multiple accounts and invites multiple users, it is an abnormal situation; if a page that should be browsed by mobile phone is visited by computer browser in large quantities, it is an abnormal situation; if a device registers multiple accounts and invites multiple users, it is not an abnormal situation (users can be randomly selected for telephone return visit);

2. Similar to Ele.me’s detection of user devices, IP addresses, mobile phone numbers, third-party accounts, payment accounts, etc. that have never appeared before;

3. Detect user behavior, such as whether a user never logs in again after registering, and whether all users invited by a channel or user are of this type (this is more applicable after data analysis and modeling);

4. Sesame, liveness detection (ordinary real-name can be stolen, liveness is difficult to crack), user address book detection to see if there are fraudulent users, abnormal mobile phone numbers (card merchants, cat pools), abnormal IPs (proxy IPs, anonymous IPs), identity impersonation, etc. This is the most difficult one, but there are ready-made access services, or SDKs and the like.

Author: Three Lessons

Source: Three Lessons