To Develop an Algorithm for Online Advertising with Your Own Resources or to Trust It to Specialists?

Companies waste tens or even hundreds of thousands of dollars as a result of doing their own development. This has to do with the particulars of the advertising business.

Online advertising management is a resource-hungry process: you need to find a relevant audience and site, and deal with traffic and setting up content display. It’s impossible to put together an audience profile right away, it’s impossible to compose other elements — it’s always an enumeration of hypotheses, models, and algorithms. This process takes time and costs money. And if there’s no team that has already worked with these kinds of mathematical systems then there are delays, or you don’t achieve the desired results.

This article discusses how companies deal with the problem of developing and analyzing hypotheses and why they decide to outsource development to OrbitSoft.

An agency from Spain makes money by selling ads. The company buys ads from platforms with the pay-per-impression model (CPM) and sells with the pay-per-action model (CPA) — when only targeted user actions are paid for, such as downloading, subscribing or registering.

The problem is there are no resources for hypothesis testing — neither a team with the necessary experience nor a ready-made software solution. One of the ways to increase income is to learn how to show ads to audience members that are more likely to perform the targeted action. This is helped by algorithms that set the criteria for displaying content.

Agency mathematicians have hypothesized a new algorithm. As planned, it should optimize traffic for smartphones and tablets so that users take more targeted actions.

A single hypothesis is obviously not enough to say with certainty whether the algorithm will work or not. Everything needs to be tested. Otherwise there is a risk of investing in a solution that doesn’t change anything.

Companies can test hypotheses using their own resources. This requires programmers, testers, and analysts. Another option is to hire an external team that specializes in checking algorithms.

The agency doesn’t have a hypothesis testing team, so it would have to be hired. In this case, testing would take about two years: first, you find specialists, immerse them in the process and features of the algorithms, develop a program for testing — and only then do you conduct tests. This decision is also affected by the budget: according to the agency, it would have to spend $ 200,000 on salaries.

Solution: Delegate Hypothesis Testing. The agency turned to OrbitSoft to test the hypothesis for the new algorithm. We already have a solution for such problems, so we went to work immediately.

Based on our Demand-Side-Platform Product (DSP), we developed technology that can test a mathematical model. It took us 1.5 months and $ 20,000 from the customer to develop it.

Result: With the help of ready-made technology, the customer tested the model on real users, in real time. It took 4 months to verify it, and during this time the company realized that the algorithm was economically unprofitable — they could not achieve the planned conversion with it.

Despite the fact that the algorithm didn’t work, the company still got their result. They discovered that their hypothesis didn’t help them target more accurately, so there was no point in investing in this algorithm. It would be better to develop and test another one.

The company also saved money. If they had tested the hypothesis using their own resources, they would have spent up to two years and up to $ 200,000. This way the project took less than six months and cost one tenth the price.

This company creates content for mobile operators, such as videos, GIFs, games, and so on. These partners are called content providers. End users buy their content through cellular operators.

The main way to attract customers is by displaying ads that convert users to customers. The provider receives a commission for each targeted action.

Paid content is games, videos, memes, and anything else that might grab users' attention. The more interesting it is and the more it meets the interests of users, the more willing they are to pay for it.

Problem: Ads are Analyzed Manually. The provider’s income depends on the ratio: the cost of advertising and the number of subscriptions that follow. Conventionally, the lower the price and the greater the number of targeted actions, the higher the earnings.

First, the customer contacted CPA networks, which buy traffic for impressions and sell for actions. The problem is that they optimized their ads mechanically: they bought traffic, manually analyzed statistics, and manually bought what was more efficient. It took a lot of time, and it was impossible to avoid mistakes.

Solution: Automate Ad Scoring. The customer wanted a data analysis tool that would help them earn more. Such a tool is called a model.

General idea: the model examines the data and, based on this, determines which ad is more profitable to buy, taking into account the price and conversion. Moreover, it works automatically, without human intervention.

The company turned to OrbitSoft to develop the model. We used our product and wrote a predictor for it. With the predictor’s help, the program understands when and what to purchase for relevant traffic.

The predictor studies the behavior and reactions of users to the ads they see. For example, what sites the user visits, at what time, what he or she looks at more often, for how long, what he or she reacts to, and when he or she closes them. For a more accurate analysis, the predictor processes information from 54 traffic sources — SSP and Ad Exchange.

Results: The provider automated the analysis of advertising and accelerated the process of ad buying, which affected the performance indicators for all work with clients' content (mobile operators).

OrbitSoft Predictor Helped Increase Conversions, and other Advertising Metrics Work with Content

This company worked with cryptocurrency and investments, and this niche was profitable. As another source of revenue, the owners chose to place advertisements for clients on cryptocurrency platforms.

Problem: It was not clear how to find a relevant audience. Customers were willing to pay for advertising, but the company didn’t understand how to build the process, i.e. where to place the advertising, under what conditions, how often to display it, and so on.

Solution: create an algorithm for finding an audience. The company didn’t waste resources developing the technology for the analysis and selection of the audience themselves, and immediately turned to professional help.

To solve the problem, we used our own product — OrbitSoft DSP, and developed an algorithm specifically applicable to the client’s situation. Thanks to this, the company received this in a short amount of time — just 4 months.

The algorithm works with 10 traffic sources, for example Google AdX, Smaato, MoPub, AOL, and OpenX. It analyzes the incoming audience and determines with what probability the user fits the desired segment. The more user profiles there are, the higher the accuracy of decision making in subsequent analyses.

Result: Thanks to the algorithm, the company gained a relevant audience base and increased its revenue. It began to buy traffic for 10 thousand dollars — this is the daily spend of advertisers. The larger the amount, the more commission the company receives. The result allowed the company to enter the top 10 advertising agencies in the field of cryptocurrencies and fintech in the countries of the former CIS in six months.