How a Marketing Director Can Build an Effective Analytics System

[ashammi228]

Content
"In 2018, MarTech expenses exceeded wage fund expenses"
“Business should not adapt to what an analyst can do”
Collecting, analyzing data and making decisions based on it is a necessity for any business. Ilya Chukhlyaev, Russia & CIS Director at OWOX BI, talks about how to properly build an analytics system in a company and what role tools play in it, and what role specialists play.


— Ilya, let’s first clarify why a marketing director needs analytics at all.

First of all, let's figure out what analytics is. It is a marketing tool that does not directly bring money to the company. It is a tool for assessing what should bring money - automated decisions based on data, contextual advertising or other marketing efforts. Profit from analytics is difficult to measure directly, and this leads to a lot of disputes. When this subject of disputes is transferred to the needs of the marketing director, dissonance arises: the marketing director needs to solve his problems, but analytics does not solve them directly, but only provides a large number of different reports telegram customer service malaysia from which you need to try to draw a conclusion.

I will tell you using the example of projects that we encounter at OWOX. The CMO (Chief Marketing Officer) has a task - to implement a marketing plan tied to sales from online advertising. What answer can analytics give to such a task? "Dear Marketing Director, here is a report on advertising channels from Google Analytics, advertising expenses only from Google Ads, online transactions that do not correspond to CRM data." With such input, it is difficult to calculate the ROAS (return on ad spend) of marketing activity.

But this is not the main problem. Let's say that marketing analytics has been around for a while, quality data is collected in the cloud (for example, in Google BigQuery), and reports for analyzing efficiency are ready... Can you imagine? The next question that the marketing manager will ask the analytics manager is: "Great, but how can I now turn this data into actions that will help find opportunities to fulfill the plan in the future?" Advanced tools are available, data is collected, but the task set by marketing is not solved.

It turns out that in order to achieve your goal - to fulfill the plan, it is important for the marketing team to create and build a function that will help draw the CMO's attention to:

- risks (why the plan may not be implemented);

- growth areas that will help to exceed this plan.

If such a function is built, then marketing will really receive, as they say now, valuable insights from data.

— That is, analytical tools alone are not enough; you need a specialist who will work with them and help the marketing director solve his problems, right?

That's right. If we agree that there is no magic "Fulfill Plan" button, then the company really does need qualified marketing analysts.

At the same time, I will note that there is such a global trend: if you take all marketing expenses, the payroll expenses have always exceeded the technology expenses. But in 2018, the expenses on MarTech exceeded the expenses on salaries in the marketing team. Both abroad and in Russia, the expenses on marketing technologies are growing. This means that more and more attention is paid to the fact that marketing analytics services can no longer be an expensive toy, but should really help the company fulfill its marketing plan, and the marketing director receive a bonus.

Don't waste your advertising budget
The call tracking and end-to-end analytics service will show the effectiveness of each source in terms of “from click to sale”.

Get a consultation

"In 2018, MarTech expenses exceeded wage fund expenses"
As a result, CMOs have two needs. First: to create an “ideal” marketing analytics system that will pay attention to risks and growth areas. Second: to find analysts who will work with this data and add context and conclusions to it.

— When talking about marketing services, do you mean your own solutions or ready-made products?

I know both successful large companies that use only ready-made products, and companies that have been working on their own solutions from the very beginning.

To prevent your own solutions from becoming a burden for your company, it is important to pay special attention to the development and support team, as well as documentation. In short: if your company has one genius of analytics and development, who wrote the data collection scripts himself and calculated your marketing even better than ready-made services can, expect that the analyst may get bored and quit. And you will spend a long time looking for a replacement and are unlikely to find a specialist who is ready to support this particular solution.

Ready-made services should also be approached wisely. To analyze 10 sales from online advertising per day, you should hardly think about paid services. Google Analytics, Yandex.Metrica or Excel will do. For such projects, you should not rack your brains over building your own DWH or the benefits of using ML algorithms.
But as the company and marketing tasks grow, the need for additional indicators, automation and speed of working with data increases. For example, in the SaaS business (IT solutions with payment for a subscription), you can start getting real value from DWH if your annual turnover exceeds $200,000.

Let's return to the task of "Finding growth areas in the marketing plan". Even if a ready-made service can collect, process, calculate, forecast and visualize data, it will most likely require adaptation to a specific business. At the very least, because the logic and key planning parameters of many businesses differ. This means that the structure of imported data will differ, as will the visualization. In addition, as we have experienced, clients do not like to learn new interfaces, but want data in the familiar one.

To summarize, our recipe, which we at OWOX use in creating marketing analytics systems and which we believe in, is as follows:

business should have full access to its data,
To draw conclusions, the data must be delivered to the interface that the decision maker is accustomed to.
All this increases trust in the data, without which its content is meaningless.

— Why is it so important to adapt the analytics system to a specific business?

Because even within retail, there is a significant difference. For example, the electronics business is conceptually different from the clothing or home goods business. This includes different purchase frequency and different emphasis on working with new and current customers.

Or, for example, many companies now place a strategic emphasis on the development of mobile applications and customer analytics. Such businesses have a completely different marketing model and methods for implementing the plan. And the metrics differ from the metrics of web applications. This fact complicates the unification and processing of data. In some cases, the evaluation of marketing efforts is also complicated.

There are many more of these features and differences than I have already listed. This is why it is so important for businesses to have direct access to their marketing and product data, rather than relying on the capabilities of a specific service and its visualization system.

— Can we say that just as the analytics system is customized for each industry and specific business, analysts can also be specialized in a certain area?

First of all, in my understanding, the areas of responsibility of analysts will be more closely divided and systematized. Do you know how companies now compose vacancies and requests "help find an analyst"? Especially where attention to this position is new. We want a "miracle" who will set up counters, and combine data, and form hypotheses for this data. And, of course, give recommendations so that conversion immediately increases, and advertising campaigns begin to pay off. He is an analyst, he can handle it! And there are many such requests.

In general, an analyst is sought based on the needs and pain points that have developed in the company at the moment. And if there are many such pain points, and there are always many, then the search for an analyst takes months and even years. And there is a high probability that the specialist found will not be able to solve all the tasks set. He will either not have enough time for this, or - competencies. For example, this specialist can be an excellent "technician", that is, he can competently set up collection, write a request and prepare data for analysis. But he is not strong in assessing the data received. Or another popular case: the company is large, there are many data users, but everything is organized in such a way that only one analyst can get this data. So it turns out that this specialist can be an excellent marketing analyst, is great at assessing the contribution of channels to income, but all day long he is engaged in being a voice interface to the data or a tool for regular fine-tuning.

It is better not to repeat such cases, they lead to a very low ROI of analytics. What should be done is to determine which areas of responsibility analytics can strengthen through data analysis and growth points. For example, in the retail business, these are marketing, product, and customer experience. If analytics as a function does not develop these areas of responsibility due to the overload of specialists on data and report preparation tasks, it is worth allocating this "technical analytics" function into a separate combat unit or transferring it to partners.

“Business should not adapt to what an analyst can do”
— Is it possible to find a specialist who can combine all these roles?

You can always find one, but it is more profitable for the company for analytics to be a working function, and not the capabilities of one specialist. Otherwise, the company will run into the limit of his "genius".

As an example, I will tell you a case when one of our clients' analyst suddenly suggested abandoning some of the working external solutions and remaking them himself, since he "started learning the R language and is interested in it, and the business will stop overpaying for external solutions." It seems like a good motive - to save money, but, in fact, the analyst is not doing his direct tasks, but developing his own interest. In return, the company will probably receive an unstable solution without support, if the analyst quits and no one is focused on the real tasks of assessing sales and marketing. Doubtful benefit. The business should not adapt to what the analyst can do.

The opposite story is when a full-fledged internal department of analytics and data is formed, which supports the tasks of growth in various areas. If it is correctly integrated into the processes, the company can be calm about the quality of the data and the resources that are engaged in assessing the development of marketing, product and other areas.

The head of such a department has a separate role - Chief Data Officer or Head of Analytics. He determines which data-based solutions should be developed and supported first. Which of them help the company earn more.

— How does the Chief Data Officer understand which data will bring profit and which will not?

This is where skills and knowledge in the field of marketing, product, and customer analytics come in handy.

For example, a case from customer experience analytics: a bank automates the selection of a customer manager based on the information it collects about the interests and actions of its customers. The conversion of this manager into closing deals with customers of a similar profile is also taken into account.

Or a simpler case from marketing analytics, when the configured scripts for importing expenses from Yandex.Direct did not transfer part of the data and the marketing manager made an incorrect conclusion about the ROI of the campaigns and made a mistake in planning.

Of course, such cases do not appear immediately, but after preliminary work has been carried out with the business customer of such solutions. As part of the development of the solution, it is important for the analyst to get answers to the following questions:

What are the strategic goals of the analytics customer?
What decisions and in what area of ​​responsibility will be made to achieve these goals?
What questions does the customer need to answer to make a decision?
What charts, reports, recommendations can help answer the questions?
What metrics and in what contexts should be used in the solution?
Here it is important to work through the tasks for analysis in this order so as not to miss important details and, as a result, give the customer a working and useful solution.

To summarize, for data to work for business, it is important:

to know what goals and decisions of the customer really depend on the data and on which ones,
take into account the amount of profit these decisions affect,
regularly monitor the quality of this data and promptly report any deviations from the norm.
And of course, it is in the CDO's interests to influence the increase in the proportion of employees who use data themselves, rather than requiring constant assistance from a team of analysts.

— Can an external partner help in building analytics?

I would like to point out two approaches in which interaction with a partner really helps analytics develop in a company.

First: the partner has significantly stronger expertise than the company currently has. For example, a business already wants to work with data or at least start collecting it, but there are no analysts in the company or they are focused on other tasks. In this case, the company should pay attention to a partner who already has experience in working with similar cases. I am talking about the company's niche. The partner's experience in the niche really matters. When the company has chosen such a partner, it is worth entrusting its tasks to them. But do not forget that it is important to strengthen the expertise on your side over time. No matter how wonderful the partner is, the solutions and data that generate income must be manageable. At the same time, the partner can continue to do useful work in which he is strong as an expert.

I would also like to clarify that the timeframe for filling vacancies for a strong analyst can be up to six months. That's a lot. And a partner is a great solution to start and not wait for value for so long.

The second approach: the partner's team is hired under a contract for a specific functionality. For example, this could be analytics implementation, GTM support, or marketing reporting. This approach works when the company needs large and qualified resources to perform such tasks, and internal resources are more valuable to focus on data analysis and interpretation.

But with any approach, it is worth building relationships with a partner that are transparent at their decision-making level. In other words, the partner must understand the company's strategic goals and objectives, its requirements for analytics, in order to work in a proactive format.

Don't miss the news
Enter your e-mail and receive the most interesting articles, studies and cases.
E-mail
Receive news

By clicking the button, you confirm that you agree to receive the newsletter.
— How much might it cost to build an analytics infrastructure from scratch?

At the annual Go Analytics conference, we had guys who talked about how to use paid products for analytics that cost hundreds of thousands of dollars. And there were guys who talked about how analytics can be built for $130 a month. Both are true. It's a question of strategic goals, time, budgets, and faith in success.


Go Analytics Conference 2019
For some projects, Google Analytics 360 is suitable, for example, because they actively use Google Marketing Platform products and need to combine advertising data with sales. Other projects have more mundane tasks and promotion volumes and are happy with Yandex.Metrica or free Google Analytics.

When answering this question, I would suggest focusing not on the cost, but on the functionality of the system and its profitability for the project. A trivial example: if a business sets the task of increasing conversion into sales for the analytics system , and the conversion in terms of additional income is less than the costs of the solution, it is unlikely that this solution will undergo a repeated procurement procedure.

There is also an example from a SaaS analytics study that says: if your annual revenue is between $200K and $1M, then it is time to collect data in a database, but it is too early to invest effort in working with ML algorithms. You simply will not have enough data about your customers.

With all this preparation, don’t forget the main thing: analytics should first and foremost help answer the questions “Where are my risks?” and “Where are my growth areas?” in executing the plan. If it doesn’t help you with the answers, then it’s just data that takes up your time. These two questions are the basis of marketing analytics, which should help the company and the CMO in particular make faster and better decisions. Then it will be profitable for the business to pay for both marketing and analytical services, as well as for the analytics team and partner services.

Conclusions:

Analytics must be built based on the stage the business is at.
It is not necessary to wait for a large amount of data to accumulate or to search for a super analyst for six months. At the initial stage, contact partners who have experience in building the necessary marketing analytics model in your field.
Together with your partner, choose a suitable model: ready-made products, modified for your business, or a system developed from scratch "for yourself". In parallel, you can hire a less competent specialist in the field of analytics, but well versed in the business context, who will delve into the infrastructure from the very beginning and then scale and improve it.