Inventale and MaximaTelecom About Audience Monetization Nuances
All Moscow residents and guests have already got used to free Wi-Fi everywhere. Not long ago it also appeared in subway, aero-express trains and airports – and this happened thanks to MaximaTelecom.
When the company only started implementing Wi-Fi network in the Moscow’s subway, it was a pioneer, as no one else took the risk. As starting from scratch is always complicated, the business model had to be changed and elaborated on-the-fly, with a trial-and-error method used from time to time. In September 2013 the company launched Wi-Fi on the first line, and at that moment it was difficult to figure out whether it would become popular, if ad campaigns would be effective and how vast the ad network would be. After a year, when free Wi-Fi was launched at all the stations and success of the project became evident, MaximaTelecom realized the importance of effective audience monetization.
In fact, MaximaTelecom were looking for the answer regarding how many users they would have and how they would be distributed across the ad network tomorrow, next week, and next month, in order to forecast network volume and plan the company’s further development.
The company’s business model is built mainly on mobile ads sales, and data uniqueness makes MaximaTelecom a desirable platform for advertisers and allows the building of effective targeting ad campaigns. What makes company’s data on users that unique?
Let’s see it in figures.
According to Moscow subway’s statistics for 2016, passenger traffic exceeds 9 million people per day, and passengers spend more than 5 hours per week on subway trains in total. On average 48% of adults use the subway daily, and this makes it nearly 5 million people. The daily audience of “Wi-Fi.ru” counts more than 1 million unique users. These are mainly Moscow citizens aged 25-34, active users of internet, digital tools and technological novelties. MaximaTelecom carefully analyze their audience and collect such information as subway lines and stations, typical routes, interests, and places of interest, devices, which are used to access Wi-Fi, etc.
MaximaTelecom is a large company that works as a complex mechanism, smoothly and efficiently. Apart from maintenance crew, most of the team work solely on network audience monetization. Three company departments process and use data on users. Commercial managers work with clients and need to know what to offer to satisfy a customer’s request; account-managers collect data on separate clients and ad campaigns, on inventory and its availability status; traffic-managers analyze total network volume and help account managers monitor clients’ ad campaigns delivery progress.
Below are examples of tasks which MaximaTelecom managers solve during their work:
- development of sales strategy,
- media-plans building,
- evaluation of remnant traffic volume and structure,
- making reports on future network volume,
- processing deals in a “booked” status,
- forecasts of ad campaigns with reach and frequency capping goals,
- check of inventory availability when closing new deals,
- ad campaigns delivery monitoring.
What are all these for? Let us take a closer look at an example of how a commercial or account-manager works. When a client comes to MaximaTelecom, they are interested in getting a result from their ad campaigns in the first place, i.e., showing their product or service to the targeted audience, so that the product would find its customer and the service would be desirable.
Imagine a coffee-shop network, wanting to attract customers that have not been to their coffee-shop for a long time, through a special promo ad campaign. The client is interested in users at the age of 25 to 45 years, living or working near certain subway stations, where these coffee-shops are located. Furthermore, they have their phone numbers, which these users left to get customer loyalty cards. The ad campaign is planned to be launched for a week period.
What will be the actions of a commercial manager? They will apply to different information sources: CRM reports, ad server’s statistics, specifying information from other managers and account managers, and internal company data. All this data is to be processed and analyzed. In case of simple targeting, the total inventory volume can be calculated in an Excel file. However for a complex targeting (with more than 2 parameters involved) and also when one needs to consider the inventory already occupied and booked, or frequency capping, the help of special tools is required, as Excel will not cope with such task. MaximaTelecom has chosen Inventale Forecasting as the service to help with this sort of challenge. The manager has just to input all the necessary parameters into the system (campaign dates, subway stations, age category, etc.) – and get the full picture on available inventory, campaigns with similar targeting types, already running or booked for future, as well as the number of sought users available. As a result, the manager has the forecast on client’s campaign delivery, sliced by any targeting parameter, in several seconds. And if the system forecast shows impressions shortage, the manager can quickly find free impressions on similar targeting slices and offer the customer an alternative variant for their advertising campaign.
Inventale Forecasting considers the behavior of each unique user in its forecasts. Just think about it: almost half of the subway passengers are not from Moscow, and they don’t use subway regularly, but from time to time. At that, we cannot exclude these users from the data, but on the contrary are to monitor and analyze their behavior.
Let us have a look at the example. Mr. Ivanov lives in Moscow, regularly travels by subway from home to the office, goes to the bar next to his office each Friday, buys food at a supermarket on weekends and so on. This data is enough for building an extrapolation model into the future. Here is another variant: Mr. Petrov travels by Moscow subway only a couple of times per month because he comes to Moscow once a month on a business trip, but he also uses subway Wi-Fi. This is a more complicated case as we have little data on Petrov’s behavior and he acts differently, nonetheless, it is still possible to forecast his behavior patterns. The more often a user uses Wi-Fi, the more predictable they become, the easier it is to build their behavior models and trip scenarios. Merging unique users’ behavior into groups, one can draw conclusions on models of their behavior in the future. Let us take a large company from FMCG sector for example, that sets its long-term ad campaigns not just with narrow targeting only, but also with frequency capping. Ads are to be shown with a different frequency to each group of users – and the forecast is to consider this and reconstruct the behavior of all the groups to show how they will act in the future and how effective the ad campaign will be in different days of its delivery.
Audience data, their grouping, behavior models, application of targeting parameters and frequency capping – all these are terabytes of data, analysis, and evaluation of which is a non-trivial and time-consuming task. Inventale has a combination of expert rules, methods of mathematical statistics, heuristics algorithms, different methods of machine-learning and multi-agent systems as its foundation. At that, the service does not use users’ personal data but works with obfuscated data. It reconstructs the behavior of users on websites, applies information about current and future campaigns, overlaps, priorities, uses different clustering and data organization tools – all to get simple figures, so urgent for a manager: where and which user will be tomorrow.
Inventale applies different forecasting methods, allows building imitation modules, suitable for specified client’s audience and aiming at solving their business-tasks. The tool combines methods in such a way so as to get a forecast of maximum accuracy for any ad network complexity and inconsistency, and any time period. Simple extrapolation does not give the necessary accuracy, as it is essential to estimate all peculiarities of a particular website and its slice and dynamics. That is why smoothing coefficients, trends, and seasonality are applied. Forecast verification is to be at the highest level possible, as it has a direct impact on the solving of business tasks - and consequently on company income and revenue.
MaximaTelecom progresses dynamically, enters new markets (Moscow public transport, St. Petersburg’s subway, aero-express trains, airports of other cities), is not afraid of innovations and is goal-oriented. The company builds a general network, consolidates passengers from different transport (from planes to subway), which makes the process of collecting information on users and using it for creating more targeted and personalized content more comfortable, for users to access the internet freely and ads to be useful and informative.
Inventale and MaximaTelecom have been working together for 3 years. Due to Inventale’s service, the company saves about 25% of the time of employees of all its three departments. Inventale processes hundreds of forecast requests from tens of MaximaTelecom users in real-time and with that the forecast accuracy reaches 90% and higher. Current “Wi-Fi.ru” monthly audience is 7 million people. 70% of MaximaTelecom’s income comes from advertising revenue, and the company does not stop on what has been achieved.
Whatever complex your audience or the business itself would be, the more you try and the more dynamically you adapt to the changing market conditions, the more actively you use new tools for solving your business tasks, the quicker you will achieve success. Audience monetization is a process that requires analysis of large volumes of data. It is important not to underestimate the complexity of this job; otherwise, it will be difficult to be successful. Take MaximaTelecom as an example - use all your capacities and move forward.
