How to set up advertising of financial services at a variety of limits

Roman Evstafiev, a leading expert on contextual advertising in Mello

Agency Mello cooperates with JSC CB "Let's go!" from June 2019. The bank has regional offices ─ Directorate. Setting advertising has become a difficult task, not only because of the specificity of the banking sector, but also for other reasons.

Casey tells about what we face and how to solve problems.

Initial data

Client: Commercial Bank "Let's go!", Which specializes in consumer lending.

Objective: To build a clear structure of advertising accounts, efficiently reallocate funds within the directorates to establish rates management system of performing the KPI.

The starting point

Most of the project data are under NDA (non-disclosure agreement about the data), so we can not tell about the work of the previous contractor and the situation prior to the start of cooperation with a team of Mello.

Let's start with the tasks that the customer has put before us:

  • Formation of the account structure and the transfer of all advertising campaigns in one account.
  • Configure analytics for effective management.
  • KPI compliance.
  • Conducting tests, the search for new ways to get traffic.
  • Scaling with low growth CPL (Cost-per-Lead ─ charges for lead).

First steps

Two areas of work

First of all, we have divided the work on the project in two large blocks:

  1. Optimization of old campaigns and tailor the work to them.
  2. Launch of the new advertising.

Further prioritized tasks on time, not to take on everything at once, and made up for the project simplified Gantt chart. Below, one embodiment is represented by the table. In the future, we edit it on the basis of incoming tasks, their own ideas and analysis.

This chart is useful for scheduling tasks on major projects and allows the customer to see the launch of the strategy.

Rules Promotion Name

Many advertisers to hide the direction of advertising from competitors, targets are encrypted in a certain set of characters, but for us it is more important than readability. The designation of previously configured and new campaigns we have to use prefixed names.

Example:

  • Keys brand ─ Brend;
  • geo ─ Volgograd Directorate Volgogradskaya_district;
  • search engine advertising campaign ─ Search;
  • the final name of the campaign ─ Search_Brend_ (Volgogradskaya_district _-_ Vse_Goroda) _ {campaign_id}.

As can be seen from the examples of the prefix, the main triggers, which are segmented advertising campaigns ─ is:

  • geo;
  • a type;
  • keys.

The result is a convenient account structure. And in order to avoid confusion, we make a cheat sheet with the old and new names of campaigns.

The launch of new advertising campaigns

Description of our work in the following tables.

Context

Type the name Description Efficiency
Search and YAN: Wedding Keys to the credit to the wedding Small quantity of leads by holding KPI
Search and YAN: Education Keys with education loans Very small amounts of leads by holding KPI
Search and YAN: Repair Keys with a loan to repair apartments Small quantity of leads by holding KPI
Search and YAN: Avtotarget RK with automatic selection of key phrases Very good number of leads, low CPL
Search and YAN: Odnoslovniki Using key "credit", there was very little minuses The average number of leads by holding KPI
YAN: LAL (Look-a-Like ─ setting advertising to an audience similar to visitors) LAL used databases of customers and users from the site Very small amounts of leads, expensive CPL
YAN: times. interest Short-term interests loans, refinancing loans, credit cards Very small amounts of leads, expensive CPL

Target

Type the name Description Efficiency
MT: the keys Download the hot springs of context The average number of leads by holding KPI
All sots.seti: Credit Standard interest ─ credit and other topics on the banks Very good number of leads, low CPL
All sots.seti: LAL Like those who were on the site and those who have left the application The average number of leads by holding KPI
MT: DMP (Data Managing Platform ─ data management platform) User segments represented by third-party data providers Very small amounts of leads, expensive CPL
VK: Application Those who have logged on through a VC on credit selection platforms A very small number of leads, low price CPL

Automating the creation, optimization and otchetnostistaryh and new advertising campaigns

To do this, we hooked C50 platform that works with Google Adwords and Yandex. The system proved itself to large-scale projects with a budget of 1 million rubles per month.

C50 uses a tried and tested strategy based on CPL. And it is not a concrete number, and price range ─ how much the customer is willing to pay for each attracted client. The platform evaluates the performance of each keyword specified CPL and based on that will automatically update every day rates. As a result, the daily budget is distributed evenly and is not overdrawn.

Additional steps for the implementation of client requests

Nashipervye steps were pretty standard and did not solve all tasks. Therefore Mello team started to develop special solutions.

The main difficulty in setting up ad bank, "Come!" ─ in each of the regional directorate of its budget, which is still divided and promotional tools. In this case, our task was ─ one account for all directorates, that at times complicates the work with optimization.

Our actions in the given conditions:

  1. Development directorates of filling the table.
  2. Creating dashboards for internal cost control.
  3. Running a multi-level pooling and implementation of additional add-ons for it.
  4. Drafting directory UTM-label for social networks.
  5. Determining allocations of formula instruments within each management.

Table filling managements

Merge into one of 15 accounts:

  • renamed advertising campaigns;
  • UTM-created label templates;
  • We checked in each Directorate types of campaigns and the cluster key.

The last step of collected information into a table directorates filling. In simplified form, it looks like this:

Red color missing campaigns. This provision to increase the coverage and possibly getting cheap loan applications.

Internal dashboards to monitor expenditure

Our customer's 15 directorates, each with its own budget and 4 advertising platforms. It was necessary to spend money evenly throughout the month and get the maximum profit with each instrument.

To solve this problem with the help of Google Analytics add-on for Google Spreadsheet create dashboards, which takes into account:

  • budgets;
  • residues;
  • average consumption.

The table has a visually intuitive trigger ─ simple rule selection of cells that signals the urgent editing budgets / bids.

Multilevel pooling and additional add-ons for it

To optimize performance, control costs and efficiency, we used the pooling of advertising campaigns. Learn more about this tool, visit our case (a reference to the case of kabel.rf).

Pooling allows you to set rates for all campaigns averaged. At the same time some of them work better, some worse. Here comes a simple and intuitive solution: where Lida cost dear, it is necessary to cut rates, and in those cases where the cheapest ─ improve.

It becomes clear that we need different pooling. The relationship built between the advertising campaign and pooling on the basis of its effectiveness.

Advertising campaign Number of clicks Lida cost Lida What do we do?
A 100 eleven 150% of the KPI We downgrade by 1 level
B 99 12 170% of the KPI 2 to lower the level of
C 102 eleven 110% of the KPI Reserve at the current level
D 120 10 100% of the KPI Reserve at the current level
F 89 14 50% of the KPI Increases by 1 level

pooling levels:

  • 2 (upper) level ─ increase, rates increased by 50%.
  • Level 1 ─ increase, the rates increased by 25%.
  • 0 level ─ rate increased by 0%.
  • -1 ─ level decrease, the rates were reduced by 40%.
  • -2 (lower) level ─ reduction, rate reduced by 60%.

we also set up and use an auxiliary event, which was exhibited rates, focusing on the effectiveness of the keys in absolute terms:

  • cut off keys that have worked for a long time and is inefficient;
  • do not cut off the keys with an average term of use and efficiency, but at the same reduced rate;
  • increased rates for phrases and words that have been effective in the past three days and brought in Lida, for example, 100 rubles.

This means that in pooling the work was on the establishment of optimum rates and dokruchivanie effectiveness of advertising campaigns, based on already from other metrics. On this project it developed several such rules, for example:

Reference UTM-label for social networks

The targeted advertising on social networks has been confusion. In order to avoid this problem was created by UTM-reference marks for the project. Since it became more convenient to construct in your Google Analytics reports on the names and monitor the situation. The handbook clusters encrypted UTM-clear labels that are added to the advertising that allows you to keep track of where they were used. Example:

allocation formula for instruments

We needed within each directorate to effectively distribute funds among advertising channels (myTarget, VK, Yandex and Google ADS), focusing on their effectiveness.

To this end, we have developed a formula for distributing funds that takes into account the general situation on the management and effectiveness of the channels, and then adds or reduces its budget.

The formula is as follows:

The difficulties that we encountered during the project  

monitoring results

It was important to aggregate statistics by channel in one place. The main instrument became Google Analytics, since that is where the customer on a par with us the monitor the results of conducting advertising campaigns and do different sections for internal reporting.

Problems with moderation

The topics of banking and financial products is very important to observe the rules of designing banners, which are very often wrongly rejected by the system. A major problem with which we are faced ─ addition to banners and USP disclaimer, the amount of which, in the opinion of moderation, was always small and unreadable, so our achievements rejected.

Traffic Analysis for geo

For internal reporting to the client needed cuts in the number of traffic lidam and expense budget-level cities and villages. We are faced with inaccurate analytical systems to determine the location of clicks and sessions. For example, according to the Yandex.Direct, a click was made by the user from Volgograd, and Google Analytics attributed the session Volga.

Result

Most of the difficulties were expected, some of them are not solved and we are systematically fighting them. On this project an excellent management, plus a competent distribution of importance and urgency of problems, so, despite the difficulties, we have managed to achieve these results:

Month The ratio of the first month of the current lidam Lida Cost of KPI
June 100% 48%
July 215% 62%
August 185% 67%
September 172% 79%
November 223% 63%
October 240% 66%
December 230% 68%
January 227% 62%
February 229% 68%
March 162% 71%
April 198% 69%
May 181% 63%

The situation on the stability of the project and understandable to the client. Due to the flexibility instrument for the lead price is acceptable, but it can change qualitatively, the main thing to do is unstressed pace.

Conclusion

The hardest part of this project ─ account management and balance between the retention of cost and effectiveness. Since we are based on the principle of hypothesis testing: run tests, reserve successful and cut off the failed ─ it allows us not to go in blind, forecasting, and work directly with the result, scale, and cultivate it. While we are cooperating with the bank, "Come!" 9 months, there are plenty of hypotheses that await implementation.

Our concepts and principles have helped to structure the approach to large projects to achieve a transparent and efficient management. A maximum immersion head of online marketing group JSC CB "Let's go!" in collaboration influenced the rapid development of the project.