Five Key Reports For Business Intelligence Project Experience LPgenerator

My name is Artem / /. In 2011, I together with his brother founded the landing platform LPgenerator. Over the five years of the project, we made a lot of mistakes and did a lot of insights about the practices of Internet marketing for b2b cloud. I will share my experiences and talk about how behavioral business Analytics can change the way you think about users, their actions and strategies of business promotion. I know firsthand that the majority (the exact number will not say, I think about 90%) cloud companies at different stages and in different niches operate without proper business intelligence, and management decisions are made not on the basis of the data, and because it wants to CEO.

Of course we were no exception. The basic mistakes that we made while working on a whim: No matter what kind of shake you. It is important that you rely on when making decisions in development and marketing. If you want to optimize costs, increase margins and make more money, it is necessary to track and analyze data.

A couple of years ago, we made a first attempt for setting up a business intelligence — its not just about checking the traffic sources in Google, the daily count of incoming leads and evaluating a basic sales funnel top-level. We need to know what is happening in the “black box” we call the user inside the system, site, product, or platform from the first session and to repeat purchases and outflow. It turned out that neither Google, nor Yandex, not suited for cloud solutions due to the fact that the counting statistics is based on sessions, and we need the data in the context of real user actions (behavior analytics) — that is, real data, and not average in the hospital, and only within simple linear funnels. After a quick search, and analysis, we found two strong serious service, sharpened by clouds — Kissmetris and Mixpanel (remember that we are talking about 2014, today such decisions even more). Introducing (then we thought I configured everything correctly) Kissmetris where consolidated information for follow-up tracking funnels and reporting, we were happy that finally received the correct behavioral Analytics and will be able to use the data to scale the business.

But the Grand plans collided with reality. Knowledge certainly is power, but simply possessing information is not enough. Data must be able to work. To be able to interpret the reports, customize the funnel, look for a narrow part of the conversion tunnel. Its one thing when you have the numbers, and quite another — when you know how to analyze, understand and apply Analytics in practice.

Then for two years we did two more rounds to Mixpanel, then back to Kissmetris, then a few times ordered a custom configuration of Google, and guess what. Thats right — it was all very cool in theory, but did not give results in practice. Trying to configure the business Analytics LPgenerator, we came to the main conclusion — we need a competent person who, using relevant tools to work with data on a constant basis. At the end of 2016, we decided to build our own subsidiary PrimeData, business intelligence to the clouds. During the building process of the correct interpretation we need behavioral data, we have identified the top 5 reports needed by the Manager and a marketer in order to make informed decisions in promoting and developing.

I think that these examples will be useful for marketing Directors and founders of startups in various stages of development. Actually sounds just right. Im willing to bet that in most cases you incorrectly believe the ROI context, media or other advertising. The focus of the following:

Advertising as a channel for attracting: Advertising as a channel involved in the conversion of the user in the first session: Advertising as a channel involved in recensione (renewals):

Some examples: The user card stores all the information about its interaction with our project. Data from different subsystems of our company (website, backend, service, support, CRM, call center and others) are transmitted to the user card, and we know his entire “history”. Data on users is stored in one place, allowing you to analyze them, building a clear tabular data. Of the card user with its properties (properties).

The values of the UTM tags, like all other properties, not pereselyaetsya and stored with time stamps. This is useful for evaluating the effectiveness of advertising campaigns of different type at different stages of the funnel and the clients life. The baseline report on the funnel (Funnel Report) segmentation based on the acquisition channels (attribution model — the first value in the selected period). We are building a funnel with different granularity and segmentation for finding bottlenecks and evaluation of advertising channels, or campaigns. Interpreted the information in the result is represented in a simple report.

On advertising spent “x” rubles, ROI funnel at the first session was “x%”, ROI funnel with extensions, “z%” — so we get a net ROI of paid acquisition channels for the specified month. Thus, we exactly understand how our budget works and what the real impact. Which channel generates attract for your business is the most qualified leads that turn into customers. The question is simple, but most marketers dont know the answer. Also included is the report about which we spoke above, but thats not all.

How are you doing with affiliate marketing. Your mailings give the effect. What is content marketing, what type of posts to your blog generate customers, how to build a proper media plan. SMM pays for itself. And affiliate program.

SEO-promotion, branded queries, and so on. And what is the ratio of all channels to attract and where to focus efforts?. These are some examples of reports that we use to evaluate ladogeneratory and clientgenerated channels: An example of a summary report on new registrations and payments. For the selected time period we see the number of new users, number of payments, conversion payments, and revenue for each channel for attracting.

If necessary, you can make a nested detail. An example of the detail of the channel Organic. Google at first point (page) log on to the website. We can see exactly which page and which posts give us new users and revenue. Part of the report on revenues (“Revenue Report”) with segmentation by channel for attracting (the attribution model, the first value for all time).

We can clearly see which traffic channels for the first time to bring users, who become customers. The report displays key metrics for SaaS each segment. Segmentation can be performed by any parameters (properties) that are recorded by the service. Summary report of affiliate (referral) program. We saw how the number of the users we attract partners.

The top partners in calling us “hot” leads, there is much to learn and apply it. Cohort analysis is a method or approach to research data. The use of cohort analysis for cloud services is especially important, as the payment of the tariff plan in most cases is not at the time of the first visit or registration, and offset at least a period of trial access. Using the calculation for holes, we just see incomplete data. First, we use cohort analysis to calculate the ROI of advertising channels.

It is important to understand that lead generation we pay today and they will be in two weeks (after the expiration of the trial access) or later. It turns out that between advertising costs and revenues are offset (window). We do cohort of new users via advertising channels or campaigns and see how they buy and how much do they give for several months. Second, cohort analysis allows you to find loyal customers and fans of the product. We break the users into cohorts according to the conditions (time of cohort or cohort by property values) and examine them.

An example of a cohort report (Cohort Report) to assess conversion “Registration → Payment” subject to pending purchases. Cohorts are formed on the first channel of attracting. To calculate the conversion take only the first purchase. An example of a cohort report (Cohort Report) to search for loyal customers — Power users (customers who pay for service for, for example, four months). Of the cohort report, you can upload a list of users from each cell.

On the same principle we find users who constantly use certain features of the service. Onboarding-the process is usually called the behavior of users in the system from which you can allocate a certain pattern to indicate that the value of the product term and the probability of conversion to paying customer is increasing rapidly. This is that same black box of your product. Onboarding is usually activation markers (action) in the first session after registration or when you first visit your site. You know, what you have is an activation action.

Or think you know. And it is the Foundation of effective Growth Hacking today. The search process of activation is similar to reverse-engineering. The data for such a reverse engineering we can find in the Kissmetrics reports and further analyze them to identify the “bottlenecks” of the funnel. Usually we see a situation where there are many ways in which users go from the moment of registration prior to purchase.

In this set of paths has one or two “mass” (which are the majority buyers) and one or two most conversion. Further, we need to lead users through the conversion path. In the end we will get an increase in conversion “Registration → Pay” and more sales. Sample report on ways users (Path Report). We examine the way users in General from check-in to purchase (for events or parameters with different level of detail) and more details for each unit the functionality of the service.

This is the first step to “decoding” real-user behavior within the service. Using reports, like People Search, we make the sample of users on the necessary conditions (scenarios service usage scenarios multi-channel funnels, scripts, shopping and so on) and derive the desired parameters (properties) for a more detailed study. How often do you do split-tests. As you know, what to test. And how do you interpret test results?.

Marketers and founders of cloud services is often “played” by the color of the buttons or labels on CTA-elements, however, such tests usually give a local change in the micro conversions, and final, call it “MicroConverter”, the purchase does not change or even getting smaller. The point is that when conducting split-tests you need to test the impact test as a “local” conversion (a micro conversion) and “global” conversion to purchase. Studying the behavior of real users and customers (see Report # 4 “Onboarding or channels activate within the system”), you will clearly see the places in your funnel, which you want to test to increase conversion to paying. No longer need to spend time and resources on a bunch of meaningless tests. We use reports on A/B tests (A/B Test Report), which allow you to see how different pages affect the individual parts and the whole funnel as a whole.

Tests can be configured in Optimizely, VWO (for which there are ready solutions for integration with Kissmetrics) or using any other tools. An example of a report A/B Test Report. The report displays the main parameters of the test. The number of participants, number of conversions, conversion, winning option and other. Conversion is calculated by user.

It is important that you can always get a list of users (with necessary parameters) that participated in the testing, for example: An example of a list group of test takers. We can add columns with different parameters and export it to CSV for further analysis. Links in the user list is clickable — when you click go to the card user. Here is a sample list of tasks for our team in the company PrimeData:

Introducing Behavior Analytics on the basis of Kissmetris. Correct what is incorrectly configured, or implemented from scratch. Providing Data-Analytics, a specialist who will be able to:

Break open the “black box”. Usually all the posts I finish with the phrase “wish you High conversion”, but today I will add more and “Do business on the basis of data” in a new 2017.

Leave a Reply