💧 🤾🏼 💃🏼 Holistic knowledge management in an IT company 👅 🤹🏽 ⛲️

“Game of Thrones” teaches us one wisdom - resources and knowledge are at war. Resource and knowledge economies may exist in parallel, but people in them live differently. In the first, all power, decisions and values are centered around the owners of the resources: management, natural, human. Power is necessary to protect resources. In the knowledge economy, the role of each person as an ecosystem unit is important. The value of his knowledge is much more than what he was born with.

CROC is a large IT company that lives in a knowledge economy. The company has been engaged in knowledge management for many years and sells the knowledge of its employees (consulting). Technologies, competencies, languages, practices - everything is changing so quickly that this is not a fashion trend, but business hygiene.

Alexey Sidorin- Head of Knowledge and Corporate Communications Management at CROC, evangelist of big data, knowledge management and the digital economy. At KnowledgeConf, Alexey presented the chronology of establishing a knowledge management system in CROC with examples and screenshots. Read the text version of his report to learn how to build a self-governing knowledge base and gamify knowledge management.

Resource economics versus knowledge economics

“Game of Thrones” teaches us one wisdom - resources and knowledge are at war. The opposition to the resource economy and the knowledge economy is many years old. They can exist in parallel, but can completely replace each other, and none of them is better than the other. But people in them live differently.

In the resource economy, all power, decisions and values are in the hands of people who own natural, human and managerial resources. In this economy, power is important to protect resources. In the knowledge economy, everyone is important as a unit of the ecosystem. The value of his knowledge is much higher than the status and resources that he inherited.

Which world is better to live in is a philosophical question. Each company (now we are not talking about the whole world) chooses for itself what economy it lives in and it is important to know. If you sell knowledge management to a “resource” company, she will not buy this idea. On the contrary, it also works - it is almost impossible to forcefully introduce knowledge management from above, which lives in a turquoise paradigm.

I noticed that knowledge management in companies follows two scenarios.

Management decides to work more efficiently, accumulate more information, overtake competitors and orders to develop a knowledge system. A concept, a technical task is being worked out, a system is being built, employees are involved in use.

Employees Show Initiative: begin to use some tools, implement processes, best practices. The company hears the need, gives more organizational and material values that allow you to develop the system.

It is difficult to say that at least one company works one hundred percent. Knowledge management is more effectively implemented when both components are present. According to CROC's experience, the formula for a successful knowledge management project in a modern fashionable IT organization is “incest”, as in “Game of Thrones”. 30% are things systematic, structured, controlled and supported by company management, and 70% are experiments, best practices, a turquoise paradigm.

Definition of “knowledge management”

We are passionate about industry knowledge management. Recently held a session for a group of enterprises that build aircraft, engines and everything else. The session roadmap is one complex and interesting task: how to develop industry-specific knowledge management so that several companies (like a conglomerate) can accumulate an asset throughout the industry and take into account competition.

The session went on three parallel tracks. There were a lot of stickers and ideas, but always all the tracks in all teams in organizations start with the definition of knowledge management. Nobody knows what it is, but everyone understands in his own way.

Each organization determines what knowledge management is.

Note: the term itself, as well as the title of the report “Holistic Knowledge Management” is just a buzzword. It reflects what is behind it (integrity, general approach), but do not quibble over words.

In CROC we give such a definition.

Of course, we do not know what knowledge management is. But we know that the company has business processes. Process efficiency is the meaning and added value of CROC . There are three evolutionary stages in these business processes.

Three stages

The first step is to find out where the process stops, because some intangible asset is required. For example, when a client calls the call center and asks a question, the process will not continue until there is an answer. At this stage, most often people who work in the field of knowledge management work. Changes are immediately noticeable here: the process is accelerated by the time that can be measured, which means calculating the efficiency and economic effect.

At the second stage , an intangible asset is created: useful information has appeared, and our task is to preserve it as much as possible. They like to solve this problem with tools, codifications, formalization of knowledge and exchange.

Third stageThe most interesting. There is no explicit strategy in it - it is a random collision of assets that may or may not appear. Someone calls them insights, unexpected insights, "apples on the head." Here the coolest, extraordinary and effective things are born.

No one knows how to cause insights - they can neither be planned nor created. But you can create an ecosystem and conditions so that they often flare up like popcorn. But in what place this happens, no one knows.

Problems

The stages have been known for a long time, and once we reflected on why they are for us. Of course, we want to be a modern company, knowledge is our asset! This was important, so we decomposed the stages into specific and pressing problems. Five years ago, we consciously formed what we are fighting with our knowledge system.

. , . , .
, , . , , , , .
. , IT-.
. — .
«» . , , .
.
Leak of knowledge outside and loss of knowledge when an employee leaves .

Solution Tools

These are our goals. We have chosen three key and important means to achieve them.

Horizontal communications : we want to save on the way where the request goes through 5 persons. If the request goes directly - this is a clear savings.

Duplication of knowledge : one person put on a hat of some knowledge, and it immediately appears in everyone else. The faster the better.

Coordination of efforts : when several people in a company carry out the same work, our task is to reduce this work by several times and to unite. Perhaps collaboration will work better.

Methodology

When all this was fixed, a methodology appeared according to which we have been working in external projects all this time. The path is not always the same, you can vary, but, in general, any project looks something like this.

Any project consists of three layers: methodological , cultural and technical . They are divided into different branches, but work as a whole. It is impossible to perceive one layer separately from another.

The first step is the top layer of AS IS . On it, we understand how our processes work.: in which the highest added value, how intangible assets are critical there. We draw attention to the corporate culture: to barriers to the exchange of knowledge, to the mentality of why people want or do not want to do something. We are also studying IT architecture - where digital units of information and knowledge are born.

Then we go to the next layer “TO BE” and develop a usage scenario - who and why will use this knowledge system (not a program). What is the user's motivation, why he will not complain about the fact that he knows nothing or does not make mistakes, but will check the knowledge base before doing anything. At this stage, requirements are made to IT systems: they should allow all of these to be used, as prescribed in the script, as conveniently as possible.

At the last stage of the launch of ACTION PLAN, a conscious methodology appears . These are scenarios that translate into best practices, and platforms where everything happens (but not systems). The main thing here is employee involvement. This means that the idea is "sold" not only to the management of the company, but also to all participants in the process.

We’ll talk about these branches.

Knowledge matrix

First of all, it is important to build a prioritization matrix - with what information we start. The most common mistake is the desire to immediately create a comprehensive knowledge base that covers all skills and competencies, will be universal, growing, and even structured. This task is impossible at the beginning of the journey.

.

For example, the matrix that we built for ourselves and customers. Here it is in a simplified form, in our document it is more decomposed.

The horizontal scale is responsible for the degree of participation of knowledge (competencies) in the key activities of the company. A vertical scale for how difficult or easy it is to buy in the market.

The upper right corner is the knowledge that we sell directly . They help to implement projects and they are difficult to buy. Therefore, they are key and it is important for us to preserve them. Below are important and necessary knowledge that is easy to buy in the market, for example, programming. Archive knowledge is in the upper left: old projects, presale activities, where we tried to sell something, but could not. Perhaps this is interesting information, perhaps not. The latter is general knowledge.

We have chosen our own strategy for each knowledge.

Key knowledge : we accumulate best practices, provide interaction so that people quickly do their work and can agree on how to use knowledge for work tasks.
Important knowledge : we develop competence centers, communities and circles.
Archival knowledge - we develop search and systematization.
We buy general knowledge in the market: they are light, not so important and are on sale. There is no need to build competence centers for non-critical skills and knowledge, for example, in time management. I think this is a mistake - it is better to focus on key knowledge.

Competency management

The next step is to understand how to relate to this knowledge . Often in projects, doubts arise about the verification of knowledge. More common not in IT and not in turquoise companies: “How to limit people who write content to the knowledge base? Suddenly someone writes that assembler is the most correct programming language? ”

This is not scary, because there are many levels of formalized processes. The key to success is a combination of horizontal levels . Consider the example of three layers of competencies.

Formalized - 100% reliable and official competencies, because someone vouched for them (company).

For example, in CROC, each employee has a project profile. It indicates the projects and presales in which the employee participated, and specialization is the industry in which they are implemented. This helps to understand that a person has implemented many projects, for example, for banks, which means he is competent enough in this area.

An analogue of the project profile is training certification. This is the easiest and most understandable way to formalize and verify knowledge and expertise in a certain field.

Semi - formalized - competencies with outsourcing or public assessment. For example, someone recommended a person’s skills on LinkedIn. We look, and there are notes “Business Skills” or “Team Player”. It’s not necessary that all the skills are real, but they give some idea.

Expert- competencies that no one confirms. For example, if someone has a blog on 3D printing, people think he is an expert because he seems to be writing smart things.

Each layer has its own role and they must exist together. It is difficult to limit yourself to one.

This was a brief theoretical overview of what we did to systematize our knowledge. Now I will show how it works in practice.

How does it work

CROC has an intranet - CROC live. It includes a knowledge base, employee profiles, a store, and everything else.

Screenshot

Each employee on the intranet has a profile . It contains all the information about the person and shows a map of his communications within the company. Data is pulled from accounting systems, something he himself describes: organizational circuits, contacts, colleagues, skills, as in LinkedIn. Other employees vote for skills. This is a crowdsourcing assessment method: each employee determines how much another is associated with the characteristics in the profile.

There is a store on the intranet. In it you can buy various bonuses, for example, an annual subscription to a fitness room. You can buy for points that are awarded for any work that is useful for the intellectual capital of the company. Points are displayed on the counter in the upper right of the screen. The point system is for the company a way to show the employee that his little things are noticed.

Communities

An intranet is a community system . For each activity in the company (project, presentation, presale, RND research), a group is immediately created.

A group is created in a few seconds, like on Facebook. We select a category, depending on this, the correct template is recommended. Add a description and select the type: closed, secret, accessible to everyone. You can also make it external - to work not only within the company, but with contractors and colleagues from other organizations.

With the help of communities, almost the entire CROC knowledge ecosystem is created.

The interface is easily configured in groups: elements move with the mouse, as in the constructor. No need to know HTML or attract IT specialists - everything is intuitively simple. When we created this system, we were inspired by the idea that all cool inventions and corporations started in garages. Our task is to give employees “ virtual garages ” in order to quickly unite for any activity.

When a new project appears, it receives all the information from other information channels: materials, ideas on the project, correspondence. Groups are integrated with office applications to work with documents without switching tabs.

All projects are saved, colleagues can later return to them. Moreover, when starting a new project, a recommendation appears to see others. You can always understand what the predecessors did, even if they no longer work in the company. This is important to us. This is convenient, because using old document management systems or correspondence in the mail is no longer so efficient.

CROC is a large IT company with a huge staff of various internal corporate systems. By clicking on each, we also get into the community on this system, which its team supports. Each internal system has a project manager. He leads a community group and issues instructions, discussions, and news.

There is no separate team that is responsible for the knowledge base.

Each leader is responsible for their group and content. He can delegate responsibilities, but responsibility on him. This way we check how verified or not verified information is in the community and whether it can be trusted. In general, everything happens according to the described scheme, except for communities marked "official", they work a little differently.

How to lure into communities

It was difficult to create, organize, and attract people to communities. I found a 2014 screenshot that shows traffic in the knowledge system. In 2014, it was only gaining momentum, but at some point a sharp rise was noticeable.

The rise is April 1 and the competition “Poke a colleague”, where everyone laid out the counterparts of colleagues. We were surprised!

But this was not just a competition, but the first time that HR joined the promotion of the system. We had a difficult long way to start, we could not decide on the owner of the process. HR initially doubted, and then incredibly supported. HR was followed by the IT department and everyone else, and the number of people in the system increased dramatically.

Since then, engagement looks like a complete picture of communications: in the ads, tempting theses are sent to the knowledge base, it leads to the audience where interested people gather. It is a material and virtual office that are strongly intertwined.

Not everything was smooth. We had problems with the sales department.

A verbatim quote from one of the high-level "salespeople."

This is a common problem in the engagement process, but it is normal. We worked in stages: we respond to the first request, to the second - we give the link: a person is less offended when you advise you to see a specific link. On the third request, we share material in the system or invite to the group.

It was a shock for us why all employees write something on Habré, but at the same time they don’t want to write on the internal network. But this is not a cultural problem. The reason is that we provide a service not of such quality as Habr. Therefore, we began to work to improve quality so that people read, listen and like.

Onboarding

We started with newbies because it’s hard to engage an old audience. Therefore, the knowledge management process begins with onboarding - we add new employees to the digital ecosystem. The onboarding process looks like this.

We thought it over in the style of an Employee Journey Map - an employee path map. The day before the exit, we begin to accompany the person. This helps the virtual assistant through the chat bot.

An example of an accompanying onboarding bot.

The bot says where to go, where you can go to the toilet, accompanies you at hiring, tells you how to find the help desk, what password.

There is a separate chat bot assistant: he will tell you the date of salary, order a pass, book a meeting room. For us, this is an innovative solution, because the technology is cheap, but ultra-useful. Botha is easy and quick to write, but it makes life easier for both old and new employees. For example, you do not need to go into the pass ordering system or book a chat room by going through a VPN in Outlook. You can ask the chatbot anything you want, and new features are constantly being added.

An example of an assistant chat bot.

Chatbot is a platform interface. We do not know how to evaluate the economic effect of this decision. But the mood of the employees shows that it really significantly changed our business processes.

There is still a mobile applicationbut it’s interesting to those with a larger office. In the digital office of the application, you can watch a map, study objects - if you click on an object, information about them is displayed.

Our employees often use only one function - "remember the parking space." The developers on each floor hung on a small bluetooth sensor for 500 rubles. When an employee leaves the parking lot, the sensor remembers which floor he left.

In addition to bots, onboarding has a mood assessment, adaptation seminars in the format of a half-hour presentation of new employees from the company's top management once a month. All together this is an interweaving of technology and people.

Online Courses

We record a lot of different videos. This is an easy way to convey information.

For example, the topic of information security is close and important to everyone, but it’s hard to get its points across to people - boring. To do this, we recorded a small series in which employees themselves play. They play different situations, for example, when a customer’s Instagram photo led to a problem.

These are funny funny videos. The questions or questionnaires after the videos are not made in the style: “Is it possible to leave your laptop unlocked at the conference” with the answers “Yes” or “No”, but creative: “You’ve lost your unlocked laptop in Bali and also have a week off.” What to do?"

For those who have not mastered the series, we developed a game. New employees are required to undergo it voluntarily, forcibly, but everyone is interested. In the game, a little man is faced with various incidents and he must remember how to behave in the office.

For cases that we no longer need, but are interesting to customers (they are related to life and security), we create virtual simulators in 3D . We even have a digital data center double. In it you can walk in special shoe covers and glasses, carry out some manipulations or evacuate in case of fire.

Creating digital content is our greatest passion. To do this, we built a special studio. With its help, we sell courses to customers and make money . Customers are happy when a video appears that popularizes their content.

Now the studio also works as a generator of internal innovation. We record videos on internal activities. Referral managers generate content about their referrals. In general, any employee can record any content . To do this, just walk to the third building and book time.

The activity is so high that we opened a school for bloggers to teach people how to record informed and high-quality videos. We have specialists who help with this.

Technology seems quite youthful and informal. But knowledge management is not only technology, but also something structured and important.

Design experience

We tried many times to collect project experience.

The experience of key projects is our main asset.

Every year we carry out 2,000 diverse projects in different regions with different technologies and different teams. In numbers, these are 10 CIS countries and 29 far abroad countries, 84 constituent entities of the Russian Federation, 2700 regular customers.

Teams quickly leave and change, it is difficult to collect experience. We introduced various motivations: virtual points, project contests. They even introduced a book of project records - the best project falls into it, and the team is beautifully photographed, they give something. We motivated people in every possible way to put together a database of projects, but it didn’t work.

It’s not motivation that works, but strict rules.

We introduced strict mechanics: until the team members write a review about the project, which will be agreed on a complex chain with the participation of lawyers, marketing and everyone else, the project will not be closed. The whole process is complex and bureaucratic, but only this way it works for critical and important knowledge, if it is necessary to ensure their integrity.

Later we changed the process a bit. Feedback is still written in a complex form, but for each project there is already an entity within the system. It reflects the life cycle: participants, expert assessments, budget, how much the project can be PR. Finally, we have created a base in which all this can be found.

Corporate Search

Insight from an IT company: Search cannot be bought out of the box.

This is what the underlying search ecosystem looks like.

This is always a complex IT project that will never end. Search is a process . Moreover, it is very incomprehensible, it is difficult to make all kinds of indices and constantly pick up new types of materials. If someone offers this product out of the box - this does not happen, it is always a service.

When we created the search, we found out a lot of interesting things. Analysis of the search gives new knowledge about the company: the most popular information, what information they seek, but cannot find, duplicate documents, information creation centers. Search is a source of information .

If you add cool corporate functions to it like an assistant, ontology or machine learning, then it becomes a key tool.

But in our case, these functions did not work. We conducted several experiments, for example, tried to search for tenders on the Internet for TK. They took technical tasks from the head of the department that he had ever won or did, fed documents to machine learning, and he found similar to government procurement sites. I don’t know why, but people used this function a little.

Perhaps we are too small for this, or our artificial intelligence is still too artificial, or perhaps not yet ready.

Feedback

When we analyzed all the information that we have, we realized that there was not enough feedback. The simplest thing you can do is collect feedback on IT systems.

Once a year we conduct a survey. In it, we specifically request feedback on those systems that the employee used (this can be seen from the logs). In addition, there is a feedback button on any page.

We added feedback to almost all processes within the company and learned a lot of interesting things. For example, many processes have no responsibility. There was no responsibility for the process of agreeing on an agreement - an important, most complex and with many cycles. No one answered: neither a lawyer, nor a financier, nor a manager. After opening, we introduced a process owner for this and many other processes.

Crocartner

This is estimated visual analytics. This is a framework that we jokingly called it CROCartner. Everyone can try this in their company.

Horizontal - system rating on a 10-point scale, vertical - the number of users. The volume of a circle is the size of the resources that are spent on it.

The result is a model. According to it, it is clear which key systems are the most important and evaluated, and where you need to change something, because a lot of money is spent, but the system receives poor ratings.

User rating is the main KPI of most processes within CROC . Finances are at the forefront of everything, but you can understand how to improve them through user ratings.

Measuring Time2Market Employees

This simple solution is an additional button to the portal or the same chat bot. When an employee knows where the toilet and the dining room are - presses a button, receives a pass - presses a button, sets everything up on a computer - presses a button. He marks each passed stage, and we see how much time all this takes.

Using this button, you can measure Time2Market employees and get feedback. Only by measuring this process, we realized that getting to work is not 3 hours, but several days. Any perfect process can be improved by adding the right feedback.

Communication Analysis

Moving in this direction we analyze communication: how people correspond, communicate in the mail or in the knowledge bases. Based on this, we build more accurate schedules, for example, of the one who works most on weekends, what the project looks like based on communications or human utilization.

The analysis gave us valuable information not only for operational decisions at the employee level (importance and value). We got a key advantage in any negotiations on changes and transformation within the company.

You can not just say: “I see that your salespeople are not helping me, teach them better,” but confirm with facts: “Over the past quarter, the production and sales department communicate less. This will result in sales dropping in 3 quarters. Let's organize a meeting where the production will once again tell the sales people what it does. ” Now the conversation is more interesting and more detailed.

New gamification

Another example of our experiments is easier and more advanced gamification. We are implementing it as a pilot on several projects, but we will launch it soon.

It works like this: after the project, the person who participated in it is offered to give 3 points to the team members. You can give one point to everyone, three to one or not give to anyone - as you like. This is a subjective survey, everyone is equal here. A person gives points, but we understand who the informal leader is.

Thanks to gamification, we understand who the manager would like to give more money in order not to lose a valuable developer. But more importantly, it is also a source of key information: we identify those who draw out specific activities, for example, networking.

From the funny - on the main page every morning a simple poll appears, for which we give 5 points. Every morning, a colleague appears on the screen and you need to specify the name or department in which he works. The mechanics are simple, but well involved - the whole company is involved in it.

What didn’t take off

Not all the projects that we developed were able to be implemented. I’ll tell you about two projects, outline of the functions that we tried.

We have developed a network of business contacts so that employees exchange them. Someone has important acquaintances, some will publish them, while others will look and offer to sell them something.

But that doesn't work. No matter how many were launched, it did not take off. It is quite difficult to extract a contact - you can’t write phones, just contact the contact owner for permission.

This is a complex process - just because people do not share their contacts. The employees have a lot of them, but even keeping business cards on the table is a huge job, not something to transfer to the system. More often, people show off contacts or publish something at random. They still publish something, but the activity is weak.

Another example - with one Russian partner, we developed a system of analytics of external sources of information . She automatically monitored all references on the Internet to us and key competitors. We studied personnel shifts to lure interesting specialists, tracked large projects, initiatives, budget allocations.

The system is extremely complex - it allows you to build graphs mentioning information. We used it for a while, but this is not our profile, we are not b2c. Now we use cloud services easier.

Ecosystem of knowledge

If I were writing a technical task for a comprehensive knowledge system within the company, there would be the following blocks.

Profiles of employees.
Landing platforms. This is the ability to quickly create pages within the company, like on Tilda or Wix. Each unit can quickly make a landing, both internal and external.
Collaborate on documents and everything else.
Multifunctional communities.
Intangible motivation.
Analytics system and search.
Feedback.
Training and virtual assistant.

The core of the CROC ecosystem is Jive. This is an American commercial system used by Apple and Google.

Jive is the heart of our automation.

The basis, first of all, is a web application server . We have many (very many) small PHP crafts that are easily installed on the server, quickly launch and solve problems without requiring the implementation of a separate system. The server is built into Jive and you can quickly launch a new function through the Jive interface. Everything integrates with internal systems , for example, CRM.

Supporting systems : LMS, Jira, Confluence. Jive has replaced most of the Confluence processes, but the latter still remains - it is used by the IT department as a Wiki.

Also Jive is a search: Elastic search, Hadoop + BI for unstructured data

Conclusion

Once upon a time, I was greatly influenced by the understanding of the company's automation model.

Horizontal - the complexity of the work, vertical volume of work. Left corner - light work that is done manually. When volume increases, automation begins. The upper right corner is corporate systems that perform the most complex and voluminous work.

In the middle, it’s not clear what. Usually these are some crafts, web services, task managers - what is called shadow IT. This is a huge security risk. Old-school IT specialists are struggling with shadow IT and are not giving anything to the company.

On the other hand, it is a huge source of innovation. If your company uses Doodle to schedule meetings, it’s easier to make a corporate counterpart so you don’t use third-party online tools. thereforeshadow IT is important . Most of our innovations have grown from such crafts, experiments, or copying external practices.

Finally, three theses.

Knowledge management is never safe . This is always a risk of information security. It looks like buying a weapon: you need a safe, the child should not be allowed to play with him, the district police officer comes home regularly, you have to look after the weapons.

Any knowledge like a weapon is a security risk.

No matter how reliable the safe is, there is always a risk that the owner will get drunk and decide to get a gun. Therefore, security must be monitored initially.

Automation for automation . We at CROC are sick of this because it has given us a lot of innovation. People swear, they say that automation for the sake of automation is terrible and inefficient. But we were fans of automation with obsessions.

Change for Change. We carried out a great transformation under the motto: "We change the jacket not because it is old, bad or cold in it, but because we want a new one." We do not update the profile on Facebook because we are old, but because there is something new. The pursuit of something new is important in IT systems and knowledge. Change any cool and cool portal for a new time every 5 years, simply because the changes will attract attention again. This is how a person works - he always wants something new.

KnowledgeConf 2020 Online 18 19 , : , . , , — -. KnowledgeConf 2020 Online.

, , Telegram- , , .

Holistic knowledge management in an IT company