How to do startup analytics: a complete selection of indicators. To analyze means to be able to process the received information.

How to start sales analytics? All analytics indicators are divided in accordance with the 2Q1D principle into 3 types: qualitative, quantitative and development indicators. 2Q1D involves the use of sales analytics methods according to a certain universal algorithm.

Quantity means analytics of sales indicators in quantitative terms. It monitors how efficiently the system works: cold calls, meetings, traffic from a website or blog, Internet marketing tools. By doing this analysis, you are learning how to maximize the entrance or “throat” of the funnel.

Quality combines a group of sales analytics methods that are aimed at measuring quality indicators. How to compose it within the framework of quality? Measure and use methods such as:

Base segmentation

in the field of business (B2C, B2B, B2G);
by size of the contract;
on the subject of the transaction

ABC XYZ - Analysis
Customer and Product Migration Study in ABC XYZ Categories

Using such sales analytics, you take care of expanding the walls of the funnel along its entire length.

Developing. It's about about business development. The goal is to create new funnels by product and channel. Only after working on quantity and quality, launch new products, connect new channels. Then rates and quality of development are estimated, standard indicators of sales are measured.

Sales Analytics: Research on ABCXYZ

Sales analytics in a company starts from the current base. It will allow you to understand which customers buy more and more often, as well as which products buy more and more often. Why do you need it? Obviously, after such a study, you will be able to identify representatives of your target audience with a high degree of accuracy, which in turn will ensure revenue growth.

The letters ABC XYZ mean:

Group A - customers/products with the largest volumes of purchases
Group B - customers / products with an average level
Group C - customers/products with a small volume of purchases

Category X - those who contact you most often, or the most requested products
Category Y - counterparties with irregular circulation, or irregularly shipped products
Category Z - those who make single purchases, or products with single and unpredictable consumption.

Sales analytics: working with the current database

Current clients should be handled by separate managers. You should not combine these duties with the work of attracting new customers. Regular ABCXYZ analysis will help control:

migration of buyers from category to category,
set personal revenue plans for clients,
plan growth according to the up-sale (same products) and cross-sale (cross-products) nomenclature,
create motivation for employees working with the current base.

It is important to encourage repeat business. This can be done in various ways, including special loyalty programs that include gifts, free seminars or master classes, discounts and promotions on the range, etc.

Sales Analytics: Customer Loyalty

Clients in group 2 are also important: their suggestions on what should be changed to improve their valuation can be the basis for reducing churn and improving future work.

Sales Analytics: Penetration

If the company has a periodic product/service sales cycle, then the analytics should include such an indicator as the share in the client. It is important to understand how many products the buyer takes from your company, and for which he turns to competitors. At the same time, it is also necessary to assess what is your share in terms of up-sale and cross-sale potential.

You can get data in three ways:

statistical (available information is collected on the market, the number of companies operating on it, the number of consumers);
survey of clients according to the developed scripts;
survey through a market research service.

For example, a company delivers 40 packs of paper to one counterparty every month. At the same time, the total volume of purchases is 100 packs. We calculate the share in the client:

40: 100 = 0.4 or 40%.

And now we need to think about how and by what means to increase this indicator: to offer a better price, convenient delivery, expand the range of stationery, etc.

Sales Analytics: Reporting System

What is sales analytics? This is primarily the setting of the reporting system. Reports for sales analytics are generated according to 2 main principles: functional and structural.

In accordance with the functional principle, sales analytics reflects what is happening in the business processes of the company as a whole.

And the structural principle implies analytics in sales by department.

Functional principle

Lead generation. To conduct sales analytics in this area, you need to follow 2 areas:

the total number of leads entering the funnel from all channels
sales performance analytics depending on the channel through which the lead entered

Lead conversion. This is the area of primary sales. The most important thing in it is an increase in the percentage of successfully closed deals as part of incoming traffic. Here we analyze 2 parameters:

quality of leads - their qualification for compliance with the portrait of the target audience;
conversion when moving from stage to stage.

Development client base. This is already advanced analytics or sales analysis by 5 indicators.

Penetration or share in the buyer
Loyalty Index (Net Promoter Score - NPS)
Customer value (Lifetime Value - LTV)
Customer Retention Rate (CRR)
Average Revenue per Customer (ARC)

Finance. Sales analytics is primarily a calculation of the profitability of activities for each of the customers. Companies often operate at a loss even with high turnover. Do you value your large but problematic counterparties? And now calculate how beneficial it is for you to cooperate with them.

Structural principle

According to this principle, reports are compiled on the indicators of 3 key divisions of the commercial direction.

Marketing - provides lead generation and supports the deal throughout its progress through the stages
Sales - directly affect revenue
Finance - monitors the profitability of the enterprise

Sales analytics: daily control of plan and fact

Daily as well as weekly reporting and monitoring is an important part of all sales analytics. The methods that are used in this case are the filling of 2 predictive and 2 control forms V .

Forecast forms

1. "Payment plan for the week." As an example of sales analytics for this form, we present the following table.

A similar table is filled in by each seller on a weekly basis in order to be able to form goals for the next week. A week is a critical period, since at least a quarter of the entire monthly plan must be completed during it. How to do sales analytics according to the “weekly payment plan”?

Create a similar form in CRM
Oblige managers to fill it out in certain time. To do this, make the appropriate changes to the motivation: add and enter non-monetary penalties for violation of the filling rules.
Upload completed forms weekly to discuss plans at a large group meeting, for example, on Monday.

2. "Payment plan for tomorrow." To conduct sales analytics in a timely and correct manner, you need to constantly monitor the progress of weekly plans through the pipeline. The form "payment plan for tomorrow" will help you with this.

Such a form is filled out at the end of the working day to plan activities for tomorrow. The manager must track the daily change on this form and check with the “weekly payment plan”

Control Forms

1. "The fact of payments for today." Planned sales analytics indicators are controlled using the "Actual payments for today" report. How to do sales analytics on a daily basis? To do this, employees fill out such a table twice a day.

Set breakpoints when salespeople enter data into this form. For example, they can do it the first time before lunch, and the second time half an hour before the end of the working day. Such control is justified, as it stimulates the staff to move towards the daily goal, and the manager will allow, intervening in time, to correct the situation.

2. Report "Board". It is generated automatically as a result of preliminary sales analytics. Example:

All data in this format is displayed on a video panel for public review.

Why do you need sales analytics on the "board"? From it, in a few seconds, you can determine how things are with the implementation of the plan.

The first column deserves special attention - "the percentage of the plan completed for the current day." You should not confuse it with an absolute indicator, but rather with an absolutely useless indicator that simply reflects the percentage of completion of the plan.

Indeed, if we imagine the situation that the manager has fulfilled approximately 70% of the forecast, and at the same time we are in the middle of the third weekly cycle. What does this tell us? We can understand whether a person copes with the tasks assigned to him or not? Obviously, we will not get answers to this question.

Therefore, the “percentage of the plan completed for the current day” is a measure of the pace. Thus, he "reports" how the plan will be fulfilled by each of the employees if he continues to work in the same rhythm.

From the above "board" it is clear that Sidorov is doing very badly, Ivanov should accelerate a lot, and only Petrov is working with almost 100 percent return.

The following formula is loaded into CRM to calculate the rate of plan execution:

Fact on this moment: (Plan for the month: total number of working days in the month x number of days worked per month) x 100

Sales Analytics: Call Accounting

Analytics on sales of goods and services is also carried out according to qualitative and quantitative indicators for calls.

How to do sales analytics: cold calls - qualitative characteristics

1. We draw up flow charts (employee development sheets) - a list of skills needed to successfully close a deal.

2. Every month we listen to 2-3 conversations of each seller and put down opposite the skills (checkpoints): "1" - applied and "0" - did not apply.

3. We evaluate each call according to the traffic light system: green - more than 80% of the checkpoints from technological map; yellow - completed 60−80; red - less than 60% completed.

If you have a yellow-red canvas in front of you, something went wrong in the department. As measures to change the situation, use: trainings, recruitment of new personnel, dismissal of individual managers.

How to do Sales Analytics: Cold Calling Quantitative Data

The more high-quality phone conversations with representatives of your target audience, the higher the revenue. Therefore, we control the number of calls:

number of incoming;
number of outgoing;
plan for the day
fact per day;
indicators for the entire department and for each employee individually.

Another important quantitative indicator is the duration of the call. The point here is not for salespeople to talk less or longer, but to find the optimal duration for a productive call.

Sales Analytics: Returning Lost Customers

Analytics is essential in the process of returning lost customers. Explain the reason, prepare the right script, set up regular work in this area - all this can be done with summary data at hand.

After talking with the "fell off" buyers, the reasons are clarified: perhaps there was some kind of conflict, the characters did not agree with the new manager, the product does not correspond to the declared quality, or they simply did not call for a long time.

Having analytics in this segment, you can draw up an action plan to: improve the quality of a product or service, offer an exclusive contract, send a nice gift, etc. Try to please the departed client!

Sales Analytics: Attracting New Customers

As in working with the current base, individual managers should be involved in attracting new customers: according to department statistics, this gives an increase in the number of customers by 2-3 times.

And in this direction, it is also necessary to conduct analytics. How and to what extent the customer base has been updated over the past six months, what is the conversion from leads to deals, what is the funnel for working with new customers - these and other thematic reports should be at hand for the manager.

Having research on new customers will also help prepare proposals that will increase the average check. For example, to offer additional products to the main product, to include kits or improved versions of the product at a higher price in the assortment, to charge a bonus for the average bill.

Sales Analytics: Intangible Motivation

Staff can be motivated by various professional competitions. For the majority of the team, it is very important not to lose face, to receive general recognition, so excitement can be used to support sales.

It is important to determine the purpose of the competition. For example, who will complete 50% of the plan faster or who will complete the plan for a month in 3 weeks, the sale of the old collection, who will sell more goods from a certain manufacturer.

You can invite the team to compete both for a completely material prize (a weekend trip to a recreation center, a coupon for a spa, dinner for two at a restaurant, etc.), and for the opportunity for the winner to receive a pennant, sit on the boss's chair, etc. .d.

Be sure to conduct analytics - which contests are most interesting to your employees in order to identify the most relevant forms for non-material motivation.

Yulia Perminova

Trainer training center Softline since 2008.

Basic tool for working with huge amount unstructured data from which you can quickly draw conclusions and not bother with manual filtering and sorting. PivotTables can be created with a few steps and can be quickly customized depending on how you want to display the results.

Useful addition. You can also create PivotCharts based on PivotTables that will automatically update when they change. This is useful if, for example, you need to regularly generate reports on the same parameters.

How to work

The initial data can be anything: data on sales, shipments, deliveries, and so on.

Open the file with the table whose data you want to analyze.
Go to Insert tab → Table → PivotTable (for macOS, on the Data tab in the Analyze group).
The Create PivotTable dialog box should appear.
Customize the display of the data that you have in the table.

Before us is a table with unstructured data. We can organize them and customize the display of the data that we have in the table. We send the “Amount of orders” to “Values”, and “Sales”, “Sale date” - to “Lines”. According to various sellers for different years the sums were immediately calculated. If necessary, you can expand every year, quarter or month - we will get more detailed information for a specific period.

The set of options will depend on the number of columns. For example, we have five columns. They just need to be positioned correctly and choose what we want to show. Let's say the amount.

You can detail it, for example, by country. We transfer "Countries".

You can see the results by sellers. Change "Country" to "Sellers". For sellers, the results will be as follows.

This geo-referenced data visualization method allows you to analyze data, find patterns that have a regional origin.

Useful addition. Coordinates do not need to be written anywhere - it is enough just to correctly indicate the geographical name in the table.

How to work

Open the file containing the table whose data you want to visualize. For example, with information on different cities and countries.
Prepare data for display on the map: "Home" → "Format as table".
Select a range of data for analysis.
On the Insert tab, there is a 3D map button.

The points on the map are our cities. But we are simply not very interested in cities - it is interesting to see information tied to these cities. For example, amounts that can be displayed through the height of the column. Hovering the cursor over the column shows the amount.

It is also quite informative pie chart on years. The size of the circle is given by the sum.

3. List of predictions

Often, seasonal patterns are observed in business processes, which must be taken into account when planning. The Forecast Sheet is the most accurate forecasting tool in Excel than all the functions that have been before and are now. It can be used to plan the activities of commercial, financial, marketing and other services.

Useful addition. To calculate the forecast, you need data for more than early periods. Forecasting accuracy depends on the amount of data by period - better than no less than a year. You require equal intervals between data points (for example, a month or an equal number of days).

How to work

Open a table with data for the period and the corresponding indicators, for example, from a year.
Highlight two rows of data.
On the Data tab, in the group, click the Forecast Sheet button.
In the Create Forecast Sheet window, select a graph or bar chart to visually represent the forecast.
Select an end date for the forecast.

In the example below, we have data for 2011, 2012 and 2013. It is important to indicate not numbers, but time periods (that is, not March 5, 2013, but March 2013).

For the forecast for 2014, you need two sets of data: dates and their corresponding indicator values. Select both rows of data.

On the Data tab, in the Forecast group, click on Forecast Sheet. In the Create Forecast Sheet window that appears, select the forecast presentation format - a graph or a histogram. In the "End forecast" field, select the end date, and then click the "Create" button. The orange line is the forecast.

4. Quick analysis

This functionality is perhaps the first step towards what can be called business analysis. It's nice that this functionality is implemented in the most user-friendly way: the desired result is achieved in just a few clicks. You don't have to count anything, you don't have to write down any formulas. It is enough to select the desired range and choose what result you want to get.

Useful addition. You can instantly create Various types charts or sparklines (micrographs right in the cell).

How to work

Open a table with data for analysis.
Select the range you want to analyze.
When a range is selected, the "Quick Analysis" button always appears at the bottom. She immediately offers to perform several possible actions with the data. For example, find the results. We can find out the amounts, they are put down below.

Quick Analysis also has several formatting options. You can see which values are larger and which are smaller in the cells of the histogram.

You can also put multi-colored icons in the cells: green - the largest values, red - the smallest.

We hope that these techniques will help speed up the work with data analysis in Microsoft Excel and quickly conquer the heights of this complex, but so useful application in terms of working with numbers.

Translation
tutorial

You need analytics.

I am quite sure of this, because today everyone needs analytics. Not only to the product team, not only to marketing or finance, but also to sales, shipping, today everyone in a startup needs analytics. Analytics helps to make all decisions, from strategic to tactical, for both managers and ordinary employees.

This post is about how to create analytics in your organization. This is not about what metrics to track (many good posts have already been written about this), but about how to make your business generate them. In practice, it turns out that the question of implementation - How do I build a business that mines data for decision making?- is much more difficult to answer.

And this answer changes all the time. The analytics ecosystem is evolving very quickly and the options you have at your disposal have changed significantly in the last 2 years. This post reflects the recommendations and experience of using data technologies in 2017.

First: Why should you listen to me?

I have worked in analytics for almost twenty years. I have seen many successful cases, but many more have failed. Early in my career, I implemented legacy BI for enterprises (eh). From 2009-2010, I built the first analytics in squarespace and raised a big round with this data. Then I became COO at Argyle Social, analytics startup social networks and then vice president of marketing RJMetrics, the leading BI platform for startups.

Now I help startup leaders implement analytics by being CEO and founder Fishtown Analytics. At Fishtown, we start working with companies after they raise their A round and help them build their analytics as they grow. To date, we have gone through the process that I will describe in this article with more than a dozen companies, including Casper, SeatGeek And Code Climate.

I will explain step by step how to do analytics at each stage of your startup. My recommendations for each stage will help answer the question: “What is the absolute minimum I can get by with?”. We are not here to build castles in the air; we need the cheapest solutions.

Let's start.

foundation stage

(From 0 to 10 employees)

At this stage, you have no resources and no time. There are a million things you could measure, but you are so immersed in the details of your business that you can actually make good decisions based on instinct. The only thing you really need to measure is your product, because product metrics are what will help you iterate quickly in this critical phase. Everything else goes into the background.

What to do

Install Google Analytics to your website using Google Tag Manager. Data won't be perfect without additional work but now is not the time to worry about that.
If you have an e-commerce business, then you still need to make sure that everything is in order with your data in Google Analytics. GA can do a great job of tracking your eCommerce events all the way from visitor to purchase, so take the time to get it right.
If you are developing software, you need to track user events. It doesn't matter which tool you use, Mixpanel and Heap are very similar and both are great. At this point, I wouldn't think too much about which events to track: just use the AutoTrack mode in Mixpanel or the default settings in Heap. When you realize that you need any events, you will find that they are already being tracked. This approach doesn't scale very well, but it will do for now.
Lead your financial statements in Quickbooks. Make predictions in Excel. If you have a subscription business, use Baremetrics for subscription metrics. If you are doing e-commerce, use your trading platform to calculate earnings. Don't get carried away.

If you are not technically savvy, you may need a programmer to help with GA and event tracking. All this setup will not take more than two hours, including reading documents. Spend your development time on this, it's worth it.

What not to do

Nothing not listed above. Don't let anyone sell you a data warehouse, a BI platform, a big consulting project, or... you get the idea. Stay focused. When you start building analytics, there are additional costs. Data changes all the time. The business logic is changing. Once you step on this path, you will no longer be able to pause your analytical project. Postpone big investments for later.

There will be many questions that you simply cannot answer yet. This is normal (for now).

Very early stage

(From 10 to 20 people)

You increase your team a little. These people need data to do their job. They may not be data experts, so you need to make sure they get the basics right.

What to do

You probably hired marketers. Make sure they are responsible for GA. Make them responsible for the purity of the data displayed in it. Have them put a UTM tag on every damn link they make. Let's make sure your subdomains are not being tracked twice. Your marketers might say they don't "dig GA". Don't listen to them. There is enough information on the Internet about GA so that if they are smart and motivated they can learn and figure it out. If they can't figure it out, fire them and find someone else (seriously).
If you have a sales department and a CRM, use built-in reporting. Make sure your people know how to use it. You should be able to calculate basic things like sales performance and conversion rates by funnel steps. Salesforce can do this out of the box. Don't export the data to Excel, generate reports in their (terrible) report builder. Even if you are uncomfortable right now, it will save you tons of time in the coming months.
You probably have several people in the support team. Most help desk systems don't have good reporting, so choose KPIs that you can easily measure in their interface.
Make sure you measure NPS. Use Wootric or Delighted .

What not to do

It's too early for the data warehouse and for SQL-based analytics - it's just taking too long. You need to spend all your time on business, not analytics, and the easiest way to do this is to use the built-in reports of the various SaaS products you already work with. In addition, you do not need to hire a full-time analyst. Now there are more important things to spend your limited funds on.

Early stage

(From 20 to 50 employees)

This is where things get interesting, and the changes over the last two years are obvious. Once you raise your A round and have 20+ employees, you will have new opportunities.

These possibilities are due to one thing: technology in analytics is rapidly improving. Infrastructure of this kind, as it is now, used to be available only large companies. Its benefits? More reliable performance, greater flexibility and more suitable platform for future growth.

This is the hardest and most important step: promising if you do it right, but painful if you do it wrong.

What to do

Set up your data infrastructure. This means choosing a data warehouse, ETL and BI tools. For data stores, consider Snowflake and Redshift (I prefer to work with Snowflake if I have a choice). Take Stitch 1 or Fivetran as your ETL tool. As for BI, look at Mode and Looker 2. In this area many, very many products; these six are the ones we return to again and again with our clients.
Get a strong head of analytics. Along the way, you will need a whole team of analytics specialists: engineers, analysts, data scientists ... But for now, you can afford (no more than) one person per staff. You need to find that special person who will add value on the first day, but who will also be able to hire a team around him as he grows. This man is hard to find- spend time looking for it. Often these people have experience in consulting or finance and often have an MBA. While this person should be willing to roll up their sleeves and get their hands dirty, focus on hiring someone who can think strategically about data and your business: they will be a critical piece of your analytics puzzle for years to come.
Consider hiring a consultant. While it's great that you've found an analytics leader, that person won't have the experience needed to bring together all the components of your technology stack or solve all the analytics challenges you'll face in your business. Mistakes made at this critical stage will cost you a lot in both time and money as you grow, so it is important to lay a solid foundation. To do this, most startups today prefer to work with consultants to help them set up the infrastructure and then build a team around it.

What not to do

If machine learning is not a core part of your product, don't hire a data scientist just yet. To build your analytics team, you need a generalist, not a narrow specialist.
In the name of all that is holy don't write your own ETL. You will spend a lot of time on development. Buy turnkey solutions from Stitch or Fivetran.
Do not use any other BI tool other than the two mentioned above. Otherwise it will turn back to you later big spending.
Don't try to get by with a more traditional database like Postgres as your data store. It's not much cheaper, and you'll spend a lot of time migrating from it later when it's exhausted. Postgres doesn't scale as well as a real data store.

middle stage

(From 50 to 150 people)

This step is potentially the most difficult. You still have a relatively small team and few resources, but you will be asked to provide increasingly sophisticated and diverse business analytics, and your work can directly impact the success or failure of the company as a whole. Nobody pressures you.

The important thing here is to move forward, making sure you continue to lay the groundwork for future stages of your growth. The decisions you make at this stage can cause you to run straight into a brick wall, if you don't think about the future.

What to do

Implement a robust SQL-based data modeling process. Your data models serve as the core business logic for your analytics and should be used in everything from BI to data science. Make sure your process allows to all users make changes to data modeling scripts, versioned and starts in transparent environment. We maintain an open source product called dbt , which is used by many growing companies to do just that.
Migrate from existing systems web analytics and event tracking Snowplow Analytics. Snowplow does everything the paid tools do, but it's an open source product. You can either host it yourself (and just pay the cost of your EC2 instances) or pay to host the event collector on Snowplow or Fivetran. If you don't make the transition at this point, you won't be able to collect much more detailed data, and get ready for some really huge bills from Segment, Heap, or Mixpanel in the near future. Once you get past this stage, paid tools can easily charge you $10,000 a month.
Develop your team thoughtfully. The core of your team should always be business analysts: people who are experts in SQL and your BI tool and spend their time working with business users to help them get data. It is incredibly important to find out what the profile of these people is, how to train and equip them. You should also hire your first data scientist at this stage. It's important to get your data infrastructure and core analytics team together before hiring experienced (and expensive) data science talent, but at some point you'll need to add those skills as well.
Start Selectively Solving Some Prediction Problems. Forecasting is more complex than just calculating quantities and amounts, but there are a few key areas that it makes sense to start diving into. If you're in SaaS, you should be working on a churn prediction model. If you're in e-commerce, you absolutely need to work on a demand forecasting model. These models may not be super complex, but they will be a big improvement over the random numbers in an Excel spreadsheet that someone in finance built.
Take the time and effort to understand marketing attribution. We could write a separate post about this, but suffice it to say that you simply cannot entrust this critical business task to a third party.

What not to do

It's easy to get carried away and start investing in a powerful data infrastructure. Do not do that. At this stage, large investments in infrastructure are still costly entertainment. Here are some tips on how to stay flexible:

Push SQL and your data warehouse hard. At this stage, you can handle anything using the processing power of your data warehouse. Buy as much storage capacity as you need - paying for servers is much cheaper than paying for people.
Add Jupyter Notebooks for data science tasks. If the data has been pre-aggregated in your storage, you won't need to do any processing in a Spark or Hadoop cluster.
Find inexpensive ways to do ETL of data for which there are no ready-made integrations. This is one of the things we love about Singer. 3
By avoiding the cost of monkey labor, you will be focused on solving real business problems.

growth stage

(From 150 to 500 employees)

This stage is about building analytic processes that scale. You need to balance getting the answers you need Today, with the implementation of analytical methods that will scale as your team continues to grow.

By the time you have 150 employees, it is likely that only a small team (3-6 people) of them will deal exclusively with analytics. By the time you have 500 employees, it could easily be 30 or more. 3-6 analysts can act quite haphazardly, sharing knowledge (and code) in an informal way. By the time you have 8+ analysts, the process will start to fall apart very quickly.

If you don't make it through this transition, you will actually work worse and worse as your team grows: It will take you longer to get useful insights and your answers will be of lower quality. It's just a non-linear increase in complexity: you will have more and more data and more and more analysts working with them. To combat this, you need processes to work together reliably.

Don't accept excuses. Doing analytics at this level is hard work and requires a talented and motivated team that is constantly coming up with something new and improving. Code review takes time and energy. Analysts are not used to checking their code. And the documentation painstaking work. You will meet resistance to these practices, especially among older members of your team who remember the "good old days". But as complexity increases, you need to evolve your processes to adapt to it.

These processes actually make analytics easier, faster, and more reliable, but their implementation is like pulling teeth. If you're serious about scaling your analytics, you'll get ahead.

You are a pioneer

I came to each of these recommendations after several years. independent work in companies and then scaling up this approach by being a consultant. Being able to work with a number of similar clients made it very clear to me How rarely do companies do this job well?.

From the translator

Too bad I only stumbled upon this post now that Tristan mentioned it in his absolutely amazing weekly analytics and data science newsletter (subscribe urgently, he picks out the juiciest of recent articles and posts on the topic).

For the last 16 months, I have actually spent exactly the changes that are described here in Skyeng. When I joined the company in October 2016, I had to assemble a data warehouse, build a data infrastructure, organize a single access to data for the entire company. Then I assembled a distributed team of SQL analysts attached to various business units, set up communication between them, code review processes and results sharing. Now we have 20 analysts, in addition to me, and I am building a decentralized management scheme for this structure.

Thanks to Tristan, now I see that I was moving in the right direction and did not step on most of the rake.

Notes

2. I have been working with Redash for the last 2 years - it is an order of magnitude cheaper than Mode and covers almost all cases, except maybe python notebooks. Looker, unfortunately, does not officially work with companies from Russia.

3. Singer is a simple open source framework from the creators of Stitch that allows you to write custom connectors to data sources in python. For example, we made our own connector to Typeform with it in order to permanently collect the results of user surveys.

4. We at Skyeng have not yet matured to the correct code review of analytics using pull requests, but I wrote a simple script that takes all new SQL queries from Redash, puts it in master, assigns a reviewer and makes a post about it in Slack. So we do not lose in speed, but we get a stably working review process after the fact in hot pursuit.

5. The book was published in 2017 in Russian under the name Analytical Culture.
From data collection to business results.

Tags: Add tags

Hello dear friends!

Very often in our verbal everyday life we use phrases like: “logical mindset” and analytical thinking. But what this kind of thinking means and what exactly the terms mean, we can not even guess.

In fact, this kind of construction of thought can be disassembled from two sides at once. Both with the theoretical part of the question, and with the practical one. If in the first case analytical thinking denotes a high ability of an individual to make decisions with the help of dry calculation, then in practice the situation is much more interesting.

Not everyone knows that it is the analytical warehouse of gray matter that implies dominance over. That is, reason completely controls emotions, and logic controls the images that are born.

This does not prevent personalities from showing themselves as world-class mathematicians or even musicians! But how to learn to analyze incoming information? In today's article, I would like to give some effective tips for pumping the skill of analytical thinking. And before that, I will throw an essay on the practical side of the above thought process.

Description of the mechanism of analytical thinking

A person is able to skillfully structure incoming information into logical blocks. It may look like separate components that form a general picture of the idea of \u200b\u200bthe problem or topic of the question;
a person is able to quickly make a qualitative analysis of the newsbreak, and then thoroughly study the headings separately;
in case of lack of arguments or facts, an individual with analytical thinking can resort to restoring the missing puzzles with the help of logical conclusions, constructive conjectures and counterarguments;
a prerequisite is to always calculate and see several ways to solve the situation at once;
evaluates the pros and cons of each of the possible outcomes of the action taken;
chooses the most optimal solution that satisfies the highest number of his requests.

Man and types of thinking

A person, depending on the circumstance that has arisen, uses a different type of thinking:

for example, thanks to the logical type, a person is able to find the relationship between the events that occur in his life and discover the sequence;
deduction has significant differences between logic. Thus, the deductive method of thinking does not compare what is happening, but independently determines the bundle of seen processes for inference;
but the analytical mindset can be described as the most advanced way to determine one of the most optimal options for solving a dilemma;
abstract thinking (creative), allows a person to generate countless amazing ideas and creative endeavors.

In addition to the successful switching between types, it is thanks to the analysis of incoming information that people of an analytical way of thinking are able to achieve high performance both in the professional field and in their personal lives.

They are less quick-tempered and rather laconic. They hide in themselves powerful qualitiesmarked by high productivity. But it is worth noting that the "science of analytics" accompanies the individual to last days. Or rather, until a person completely ceases to exist.

Developing Opportunities

Who needs an analytical mindset, you ask? It is useful to sellers, and artists, and physicists, at the ready with bloggers. And all because with its help you can see the success and effectiveness of the work performed.

Oddly enough, it is not difficult to develop the skill of thinking analytically in children. To do this, they will need to systematically attend lectures in mathematics and simply attend classes. Plus, pay attention to technical background and directions.

But with adults, things are much more complicated. Now I want to present to you a few effective ways develop the necessary superpowers.

1. Workout or food for thought

Chess and mathematics

Analytical games are a great workout for the mind. So, chess and mahjong are excellently recommended. During the lesson, you can feel the pleasure and the real pumping of the gray matter.

You have to independently develop a strategy, monitor the enemy and calculate your moves in advance. Since the development of logic is directly related to analytical thinking, I strongly recommend that you perform all kinds of computational operations in your mind.

Computer games

And here computer games helpful as ever. Of course, this type of activity is designed for very lazy people, but nevertheless, quests and strategies perfectly develop analytical skills.

You have to quickly respond to situations, calculate the risks and opportunities, and be patient for an in-depth analysis of the situation.

Own program

In this type of training, everyone is his own master. You can personally choose the topic and flow of information for comparing arguments and facts. Perhaps you will like the study of scientific programs or magazines, familiarization with complex literature for a thorough construction of a logical chain.

Analytical articles on politics, economics and cybernetics may be suitable. Also, you can improve the skill of determining the main from the secondary. I mean, right.

2. Constructive criticism

To get comfortable with analytical thinking, you need to get used to challenging any news that comes in. Doubt everything! I advise you to act as an avid debater. This will help you learn to ask logical and reasoned questions first to yourself, and later to the state, society and the framework.

I propose to pay attention to the detailed consideration of absolutely opposite points of view. As you begin to try to combine them into one continuous layer of material, simultaneously developing each of the hypotheses, you will be able to increase your level of tolerance.

3. Train yourself to plan

Be sure to plan your life ahead. Create a calendar that clearly separates long-term perspectives and goals from short-term ones. After passing through each of the completed stages, analyze the results to derive general adjustments.

It is worth highlighting key events and important dates for accomplishment in bright colors. Thanks to this way of life planning, you develop and improve not only analytical thinking, but also your activities as a whole.

4. Communication and organicity

Be sure to remember about training the ability to think analytically at the time of communication with people. Before speaking, try to calculate in your mind possible options the interlocutor's response or the course of his thoughts.

This perfectly trains attentiveness and involvement in the conversation. Also, the technique is very useful in the event of a heated dispute.

And at the same time, do not focus on the development of one of the hemispheres. Man is a multifaceted and harmonious being. And its success depends only on the versatile development of professional and personal skills, the level of intelligence, communication skills and symbiosis of modes of thought. Just!

On this point!

Subscribe to updates, a lot of amazing topics and discoveries await you ahead! In the comments, share games to develop analytical thinking or interesting tasks for logic!

See you on the blog, bye bye!