Get best of the week

Product analyst: what does it do, how much does it make, what benefits does the business bring

Product Analyst - the bridge between business and data. He works hand in hand with the product manager and helps the product team make the right decisions. Netis author Denis Vikharev tells what product analytics is, why product analysts are valued in business, who can become a product analyst, about his tasks, salary, and tools.

What is product analytics


Product analytics allows you to see how users interact with the product. Conventionally, two tasks of product analytics can be distinguished: data collection and their interpretation.

First, the product analyst collects a bunch of numbers from different sources: which buttons users click, how often they use the product, which features of the product’s product are popular, and which ones don’t. These measurements show what happens to the product, but do not explain why.

In the second step, the analyst pulls out insights from the numbers that explain user behavior. Thanks to this, the product team understands which product it has made and where to move on.
Product analytics helps the team understand “who” did “what, when and where.” And what to do next with all this.

Why product analysts are valued in business


A product analyst tracks user events within the product, translates the meaning of the numbers into the language of the business, and provides operational recommendations for resolving the problem. We identified four tasks that a business solves with the hands of a product analyst.

Keep users in the product


Money can provide companies with explosive growth and bring a lot of new customers, and product analytics help preserve existing users by knowing their behavior, working with the problems they face and the value they receive in the product.

A new user always costs the company more than the existing one, so it is beneficial for companies to invest resources in product analytics.
Opinion of venture investor Tomasz Tunguz : “On the one hand, growth helps to raise the round of investment and shows the demand for the product. On the other hand, the outflow of customers raises questions about the conformity of the product to the market.

"Stimulating business growth on a product that does not quite meet the requirements of the market can lead to a company raising millions of dollars and needing to" roll out "."

To make friends a product and a market


It is impossible to earn user loyalty without understanding the core value of the product, which guarantees product-market fit (literally, "conformity of the product to the market"). To find the same “Aha Moment” you need to know what actions separate loyal customers from lost ones.
“Aha Moment” is the key to growth, the moment when the user understands the value of the product. Finding it will help the right conclusions based on customer data. For Facebook, this was the  achievement of the goal of “7 friends in 10 days . ”
Knowing which product metrics are relevant to the user is easier to match the market. Snowplow's Anthony Mandelli recalls how Airbnb founder Joe Gebbia,  in the podcast “How I Built It,”  talked about the pattern that I saw using data: landlords were unable to rent an apartment for a long time because they didn't know how to make attractive photos. Then Airbnb took over the solution of the problem and increased the company's revenue at times.

Solving the problem with the quality of photos increased Airbnb's revenue by several times. Source: Airbnb website

Outperform competitors


McKinsey research  shows that using data and analytics wisely enables companies to grow on a massive scale. Thanks to this, the gap between industry leaders and lagging players is constantly widening.

At the time of the study, respondents from market leading companies said that their data initiatives and analytics brought them at least 20% profit in three years.

Spot work with analytics does not have the desired effect. To become a leader, you need to create a long-term strategy for working with data.

Improve user experience


A product team can modify a product blindly. But without analyzing the results, one cannot be sure what specifically led to success or failure. Product analytics examines user behavior data in real time. This helps the team rethink the vision of the product during the next iteration cycle and take the necessary actions.

Who Can Become a Product Analyst


The company "Normal Research" found that most often people purposefully go into the profession "from scratch", and some experts come from development and marketing.

Data  from the  2019 report on the analytics market The

profession of a product analyst may be interesting for product managers who already know how to work with the value of a product, but want to develop expertise in analysis: understanding product growth points, predicting its development.

To qualify for the position of product analyst, you will have to recall statistics and mathematics.

Question to the analyst: with which background did you come into the profession?



Tatyana Chadaeva , Senior Analyst Expert at Beeline.

By education, I am an international economist with a good knowledge of mathematics and statistics. At the university she became interested in social activities and went to HR. But in the end I found myself in marketing and product analytics and I am very glad about it.


Vladislav Prishchepov , ex-analyst at Yandex, product manager at AppMetrica

Prior to his first work as an analyst, he managed to work as a technical writer and developer (JS). The experience and view of the developer often helped me in my work as an analyst.


Vsevolod Mironovich , head of analytics group at SberMarketing

Once I worked as a project engineer, led projects on development and advertising in the studio and agency. When I switched to the client side in a financial organization, they also forced me to report on the effectiveness of the launched projects: count money, plan resources, protect cases for launching new products, optimizing and promoting current ones. It was then that the need made me delve into analytics.

First, after reading the articles, I set up basic tracking and put it in dashboards, just to understand what was going on and who in general all these people were on the site. And then I began to think how to influence all this, how to increase conversion, earn more, and as a result, knowledge from the Internet, backed up by real experience and full of bumps, was enough to get a job as a leading analyst in another company.

What does a product analyst do?


A product analyst analyzes the state of a product and helps to develop it: it ensures that product metrics are not sagging and product decisions are successful.

We did not find uniform rules for organizing data, setting goals, and conducting tests for product analysts; each company has its own. The life cycle of the product that you have to work with matters: in the newly launched startup, the analyst will be asked to tidy the data collection system, and in the mature one, they will find growth points and consider competitors.

Therefore, we analyzed dozens of vacancies and compiled a  list of tasks that the analyst may face . The review included not only IT companies, but also the “traditional business”: air carriers, mass market and logistics companies.


Find product growth points and bottlenecks


It is difficult for a product team to rely on data on how many times they clicked on a button; they do not explain the reason for human behavior. Therefore, the team goes to the analyst, who draws conclusions based on the data, finds patterns and anomalies in the product.

Case of the company Devtodev



How to find a bottleneck. Gamedev example

Develop reports and product monitoring metrics


Dashboards show teams and leaders key product metrics, dependencies, and trends. The analyst decides which reports and metrics should be displayed on the dashboard so that they do not distract from the main thing and help to make managerial decisions. There is no universal set of metrics that the team needs, they are selected depending on the goals of the business and the type of product.
Product Manager Sergey Tikhomirov correlates a set of metrics  with the product life cycle . And the AppMetrica product Vladislav Prishchepov advises starting from the goal and focusing  on the key metric of the product . So, for a food delivery application, this is “time to eat”: the time from completing an order to receiving it.

Validate team issues and solutions


The analyst “digs” quantitative data to test team hypotheses and correctly prioritize tasks. If the team identified a problem during an in-depth interview, the analyst can confirm it or refute it. For example, having analyzed hundreds of conversations of call center managers with customers by keywords using speech analytics tools.

Conduct A / B tests


Working hypotheses are tested on the control segment of users. The analyst makes sure that the test is not influenced by external and internal factors: holidays, weather outside, calling customers by call center managers - all this can distort performance.

The test result should be statistically significant - according to Appsumo service  this happens only in 12.5% ​​of cases . If the control segment confirms the hypothesis, it is scaled. A separate problem is to make a decision after tests in b2b with little traffic.

Test and scale hypotheses


Testing a hypothesis is conditionally divided into four stages: 1) we are looking for a metric that we want to influence; 2) conduct a study; 3) analyze the feedback; 4) kill the hypothesis or scale. The analyst works with the product team at every stage, answers the questions “Why did this happen” and “What to do about it”, saves the team from supporting unpopular decisions. The main value of the analyst’s work is in interpreting the results.

Case of Avito company


Avito changed the interface of the product card and conducted an A / B test. He showed that in the control group they started to click on the "Write" button less (badly), but the conversion of the first messages increased (good).

The test results are negative, it seems that it’s bad.


But if you look, it’s better.

Analyze data


A large pool of tasks for product analysts lies in the field of hard skills: he understands the basics of statistics and speaks programming languages. This helps the analyst collect and process data, evaluate its quality and look for patterns.

Product analytics work - communication with people and work with professional analytics tools

Question to the analyst: the key tasks that you perform?



Tatyana Chadaeva , Beeline Senior Analyst Expert

Working with new users (not clients):

  • the effectiveness of various product sales channels;
  • construction of sales funnels, analysis of user paths, their optimization;
  • A / B testing.

Work with current customers:

  • building a user profile and calculating basic metrics (for different products): LT, LTV, MAU \ DAU, Retention, Churn, ARPU, traffic consumption;
  • search for ways to monetize a mobile application, increase revenue by products, customer retention, analysis of the effectiveness of marketing campaigns.

A lot of time is spent on:

  • setting up periodic reporting in BI systems;
  • marking up events on the site and, in general, setting up data collection from various sources;
  • building end-to-end analytics, data marts on a cluster.


Vladislav Prishchepov , an ex-analyst at Yandex, a product manager at AppMetrica

What he just did not do, but most often looked for growth points in data and numbers and highlighted weaknesses and strengths.


Vsevolod Mironovich , head of the analytics group at SberMarketing.

Conditionally, the analyst’s work can be divided into three stages: we collect data, analyze, act, and so on. It is difficult to understand all three well, because there is a huge variety of technologies around. Every time you shut up, someone makes the first commit of the next JS framework, which will reduce your hair on your head, or launch a new store of connectors to Google Analytics.

Therefore, often colleagues begin to specialize in something. Some people like statistical research more, even if the product still does it their own way, others build automated reporting systems, even if in the end they just need a screenshot for a preza. I, being the head of the department, have the opportunity to collect honey and see the results of all, avoiding mistakes. I have the opportunity, but for some reason I do not use it and go to the fields to fill my cones, it’s more interesting.

Product Analytics Tools



Python  (analogue: R, Java)

A programming language with simple syntax, a large number of libraries and a developed community that will help if needed.

Suitable for processing large amounts of data that Excel cannot handle. It allows you to visualize data, automate tasks for information analysis, build models for predicting customer outflows, and perform clustering.


Google Analytics  (analog: Yandex.Metrica, Heap) A

free tool for web analytics. It will show the sources of traffic and user actions on the site, the number of visitors, views, conversion, a report on customized events, will help to conduct a cohort analysis.


Tableau  (analogue: Power BI, QlikSense, Looker)

A platform for analyzing and visualizing data with a clear interface. It will help to build effective graphics, combine data from different sources. Works with MS Excel, MySQL, SQL, Google BigQuery, Microsoft Azure. It is possible to configure automatic updating and distribution of reports, send them by e-mail, publish a link on the server and access the report by reference.


Mixpanel  (analogue: Amplitude, Flurry, KissMetrics)

System for analytics and analysis in real time. It helps to understand what users do after registration. Allows you to build a funnel with the conditions for each event in it, send pushes, conduct A / B testing.


SQL

A tool for working with databases within the product ecosystem. With it, the analyst will receive, process and compose the necessary data without the developer. You can create reports with dynamic periods, join tables, cut off values ​​according to the necessary criteria.

Question to the analyst: three tools without which your working day does not pass?



Tatyana Chadaeva , a senior expert on analytics in Beeline

SQL, Excel, Google Analytics (+ Qlick Sense or any other BI-system + GTM, without them, too, nowhere).




Vladislav Prishchepov , ex-analyst at Yandex, product manager at AppMetrica

Google spreadsheets, Dropbox paper, analytics system / data warehouse (each product we worked had different analytics systems and data warehouses).


Vsevolod Mironovich , head of the analytics group at SberMarketing

SQL. Data is usually stored in databases, and as a rule, interaction with them takes place in this language, so without it the analyst will go nowhere. In my case, most of the data is in BigQuery.

VS Code. For the data to be in the database, you must first put it there. Sometimes for this you need to write a script in some language, which will get the data using the API of the advertising office or analytical system and send it to the destination. Coding is also useful in order to link, process, aggregate data along the way, and, in general, conduct a full-fledged study and visualize the results.

I’m just used to VS Code, because I write a lot on javascript in my free time. For work, I mainly use Python, because it has a bunch of ready-made solutions and convenient mechanisms with cells. To be in the subject, I tried to write in R, but anything at all, if only it is SUMMER - by virtue of the profession with Excel, I do not have much, as with logic and numbers.

Salary of product analysts and demand for them


A global study  by MarketsandMarkets consulting agency shows that from 2019 to 2024, the global market for product analytics will double.

Drivers are the growing use of big data and the need for companies to produce competitive products.

MarketsandMarkets: product analytics market will double in five years

Demand for analysts will also grow in traditional business sectors. For example, in retail, which transforms stores in the area into retail-tech. Retailers are interested in services for tracking customer behavior: to prevent theft, to place goods on shelves, to target ads.

X5 Retail Group  About Retailers Business Needs

Research Normal Research shows that a product analyst receives 134,000 rubles in his first year and 274,000 rubles after three years at the company.

Screenshot  from the  2019 analyst market report

At the time of writing the article on HeadHunter, there were 1,000 vacancies for Product Analyst and almost 5,000 for Product Analyst.



Product analytics articles, channels, and videos


Channels, Blogs 


  1. All About A / B Tests  - A / B Tests .
  2. Product Science  - Anton Martsen shares material on product and business strategy, metrics, analytics, applied Data Science and user research. The author digs deep into each topic in detail to convey the very essence of different methods and approaches.
  3. - — 33 000 , .
  4. Burger Data — c, - «» .
  5. Make Sense podcast — Make Sense. , — , , , .
  6. BigQuery Insights — SQL- MacPaw.com.
  7. No Flame No Game — .
  8. Krasinsky: growth, marketing & product, analytics — , -, .
  9. Datalytics — -, Python.
  10. Close2Sense — , .
  11.  — , .
  12. Grow Horse — Growth Management, , ( ).

 


  1. , AppCraft. -
  2. , Skyeng. 
  3. , Wrike. 
  4. , . 
  5. , AGIMA.  :
  6. , Retentioneering. 
  7. , Rambler. 
  8. , Ultimate Guitar. 
  9. , , « ».  R
  10. , Creative Mobile.  , , Excel 6
  11. , Devtodev 
  12. , CPO FunCorp.  iFunny

: (, , ) , ?



Tatyana Chadaeva , senior analyst expert at Beeline

I would advise you to start learning programming languages ​​right away, at least SQL. When working with big data, you cannot do without it. Good trainers:  one  and  two .

Also read a cool  article  about how managers see the ideal analyst.

Personally, it was very useful for me and helped to understand that customers expect from me not just beautiful reports, but useful insights, conclusions and, as a result, that I will know and understand the product no worse (or maybe better) than the product manager.

It’s very useful to have a good understanding of statistics, here is a good and  detailed course on Stepik , I would like to take it earlier.


Vladislav Prishchepov, ex-analyst at Yandex, product manager at AppMetrica

It is difficult to name three things that would help me. I would advise something else: more often communicate with fellow analysts from other companies, ask what tasks and how they solve, if possible, look at how they formulate and deliver conclusions.


Vsevolod Mironovich , head of analytics group at SberMarketing

  • The podcast “How Games Make” might have changed his mind then.
  • Any suitable ML course would have accelerated to 300k per second by today.
  • Something  about burgers ;-)


From the editors of Netology


If you look closely at the profession of food analatics, we invite you to study the programs of our courses:

  • « : » « ». .
  • « PRO» « Data Science». , - Python Tableau, .

More articles:


All Articles