Do you know all the features of compiling a list of requests for placing contextual advertising? How to choose the right queries for promotion? Examples of transactional requests.
Goal 1. Selling specific products
When selecting words to search for when solving the first problem, we rely on the presented model range of products on the client’s website or on the price list.
When selecting keywords, it is very important to use a structure for dividing phrases into groups. The enumeration of words occurs from words with narrow meaning to broader ones. Consider the online store of Sony VAIO laptops. First, you need to go through queries with specific product models, then – series, then – all words with the Sony VAIO brand in different spellings (English, Russian, and do not forget about the incorrect layout and incorrect brand spellings). Then you need to take broader queries with the words laptop, notebook and others. Such queries should be divided by type: thin, light, small, powerful, etc. Also make groups of words with selling additives (buy, purchase, sale, choice, etc.).
1) Selection of models Often, when we are looking for a specific product and are interested in purchasing it, we immediately request a specific product model in the search engine, for example, VGN-AW11ZR, then the keywords for the Sony VAIO laptop model range may look like this:
2) Selection of episodes
If a potential buyer has not decided on a specific laptop model, then he can request a series (Sony VAIO VGN), in this case the requests may look like this:
We make sure to “minus” all the words that were searched in the model queries. This is done so that each request has its own ad, which will increase the click-through rate (CTR) of the ad and thereby reduce the rate. In our example, from queries with the keyword "vgn" all model queries subtracted “–aw11zr –cr31sr”, since they were taken earlier and are cross-referenced with respect to queries "vgn vaio laptop", "vgn vaio", "vgn laptop" etc. Let's take a closer look. We have already taken the model request "vgn aw11zr", then we take a request with the VGN series, if from the request « vgn vaio laptop" we will not subtract the cross request with it "aw11zr", then to the query entered by the user "vaio vgn aw11zr laptop" The ad will appear either upon request "vgn aw11zr", or "vgn vaio laptop". This will depend on the product of the bid and the CTR (click-through rate on the ad), which is incorrect, since for a particular model it is better to show an ad with a link to the product that the user is requesting. Thanks to this, the likelihood of conversion from clicks to sales increases. For a better understanding, we will illustrate this example. If we represent queries as sets, then the query " laptop vaio vgn aw11zr" will be the intersection of two sets: "vgn aw11zr" And "vgn vaio laptop", since it combines both of these queries.
Rice. 1 The case when from a request "vgn vaio laptop" We we won't subtract request "aw11zr"
Rice. 2 The case when from the request "vgn vaio laptop" We subtract request "aw11zr"
When taking into account cross-requests, it is important to build a selection logic: from requests with a broader concept (meaning), narrower ones should be subtracted. If requests have the same priority, then it does not matter which request we subtract from which, so we don’t have to subtract at all.
Further, if the client’s budget allows, then the list of requests can be expanded with keywords such as: "sony notebook", "sony laptop", not forgetting to take into account cross-words and negative keywords (garbage). You can also take such broad words as "laptop", "notebook" etc. But it is advisable to break them down into smaller requests, which will be much more efficient. For example, it is worth adding selling additives to broad queries: store, buy, sale, choice and many others.
Features of compiling a list of requests for the Begun system
- Removing a keyword phrase if a broader phrase exists in our selection. For example, for Yandex.Direct the following queries were taken: "vgn cr31sr +l" And "vgn cr31sr -l". For Runner request "vgn cr31sr +l" we delete and leave only the request "vgn cr31sr" and be sure to remove “l” from negative keywords.
- Expanding a key phrase to a broader one when we don’t have this phrase in the selection. Most often, this is useful if in Yandex.Direct it is important for us to use a preposition in a key phrase, but on Begun we cannot use prepositions, but we cannot miss out on potential clients upon request. For example, "lcd monitor" should be expanded to a phrase "monitor". In this case, the meaning of the query will become wider, so it will need to be additionally cleaned of new garbage words.
In our case, the list of requests for models, series and brands for the Runner will look like this: 
popular publications
newspaper internet
Examples of keywords reflecting the essence of the news:
Obama
AvtoVAZ
Michael Jackson
political crisis
Kennedy
A correctly formed list of search queries is one of the most important points for website promotion. Errors at this stage will lead to the fact that potential customers or site visitors simply will not find it by turning to a search engine. Each resource is promoted using certain words, which together make up a list of queries, or in other words, the semantic core of the site. All words that a user might enter into a search box on a given topic must be included in this list.
In general, the selection of key queries is the very first stage of website promotion and the implementation of this stage should be treated as responsibly as possible. In general, SEO and search engine promotion is painstaking work. It doesn’t happen that you can spend a certain amount of money on links and get top results for a specific request. To get to the Top you need to complete a number of steps, and each step must be approached with maximum responsibility and diligence.
Compiling a semantic core is an important basic stage that precedes internal website optimization. It consists of several steps that are not complicated since auxiliary services are involved in the process.
First, you need to assess the popularity of the requested words. Each major search engine has its own statistics on the number of most frequent user queries. To view the demand for all the words you are interested in or to select the most suitable ones, you should look at the following services: Yandex - wordstat.yandex.ru, Google - adwords.google.com and others. These services will show the total number of people who entered a particular word into the search bar. In reality, these are potential site visitors, future buyers and customers. In addition to this data, the services provide data by region and data characterizing “seasonal” requests: by time of year, months, holidays.
If doubts arise or the words themselves are not enough, search tips will help. Many users eagerly use hints, as they perfectly reflect all possible combinations of words. For those who are professionally involved in website promotion, search suggestions are a very useful tool that allows you to assess interest in various topics or select new words for the semantic core.
There is a good video on selecting key queries, I recommend watching it.
When promoting a website, it is customary to divide all requests into high-frequency (HF), mid-frequency (MF) and low-frequency (LF). Regarding the promotion of low frequency queries, I recommend it. There is no clear gradation of which request is considered high-frequency or low-frequency, since there is a concept of competition. You can understand which request belongs to a particular group using the following examples: “news”, “construction”, “windows” - high-frequency, “documentation”, “vouchers for children”, “house projects” - mid-frequency, and “Japanese umbrellas” ", "choice of apartment", "visas" - low-frequency.
In order to better understand the topic of assessing the competitiveness of key queries, I recommend watching the following video.
Competitiveness depends on demand for the query of interest, the quantity and quality of resources in search results and is directly related to the commercial component. For comparison: “tours to Turkey” is a commercial request with high demand, and “autumn mood” is a non-commercial request, but their demand is approximately the same.
In most cases, to promote a website, a combination of all types of requests is used, and in proportions calculated by the SEO specialist. The proportions, in turn, depend on the goal that is set when promoting this site: increasing the number of visitors or purchases, increasing the memorability of the resource, and so on.
Having selected key phrases for all types of queries, taking into account their competitiveness and having calculated the proportions for all groups of queries, you can begin to compile. An adequately compiled list of frequent requests for a given topic and purpose is the key to successful website promotion and promotion, since even a small word or phrase is of key importance for search results, and therefore for the commercial success of the project.
Regarding compiling a semantic core, I advise you to watch the following video lesson.
The selection of search queries is the first and one of the most important stages that must be taken into account at the stage of creating the site itself: this list will need to be kept in mind when writing each article or material on the site. You will understand why this is so from a series of subsequent articles on promotion, but for now take it for granted. Moreover, the creation of this list is an integral part of such a stage of promotion as compiling the semantic core of the site, the creation of which I will discuss in the next article.
Where to start? And you need to start by putting yourself in the shoes of a user who has something to do with the topic of your site (blog, online store, information resource, etc.). Each user, as a rule, turns to the Internet in order to solve some problem. To do this, he enters his questions or problem statements into the search engine. At the same time, his questions can be quite specific: “how to choose a projector”, “how to reinstall windows”, “buy a video card in Novosibirsk”, etc. (i.e. it is quite clear what information the user wants to receive), and general and incomprehensible neither for a robot nor even for a person: “computer” (what does he want? Find out what it is? What types are there? Where to buy? Reviews? ), “plastic windows” (it is also unclear what exactly a person wants). As a result, users’ tasks may be different, and we need to understand exactly what problems we can help him with and what questions will he ask? To do this, let's select...
Types of search queries
Problem Definition
The user is aware of his problem and directly formulates it in a search query: “why the flash card doesn’t work”, “how to make a radio-controlled helicopter”, etc. - general queries, but they sound specific. As a result, you need to consider possible questions, and if your site has answers to these questions, then include them in the target list.
Note: You don’t need to include everything on the topic in your list. Include only what you actually have or will (plan) have on your site in terms of answering the question. If you yourself understand that the user probably will not find a solution to the problem for the selected request on your site, then you do not need to include it! This applies not only to this type, but in general to all queries that you select for your site.
Selecting a product or service to solve a problem
Let's say the user posed a problem in the nature of choosing a product or service at the previous step: “how to build your own house.” I read about it. Now he moves on to the next step: choosing specific types of building materials: “types of bricks”, “what flooring to choose”, “how to choose a used car”, etc.
This type of request is suitable for you if you have any review materials/articles, comparisons, news on the topic of the issue.
Search for reviews, discussions, tests
The user has selected any type or several types of goods/services or even specific models. Now he wants to make sure that his choice is correct or to choose the best option from the list he selected. To do this, he looks for reviews, opinions, different types of comparisons, tests, discussions. Of course, not all users and not for all products/services look for reviews and read discussions, but still the mass of them is quite large and it is impossible not to take this category of user needs into account.
This type of request should be taken into account and included in the list if you yourself conduct such reviews, collect reviews or arrange discussions, or if you at least describe these models / types of services and link to places where you can see reviews / discussions of this product / services.
Search for a specific model
The user's choice can be complicated by the set of types of goods he needs with different sets of characteristics and prices. This includes phrases like: “high-quality website creation services”, “free hosting”, etc. Next, the user has chosen the type of product or service, read a lot of reviews and now wants to read more information about a specific model. That is, it writes in the search engine something like: “Canon 600D camera”, “Epson T50 photo printer”, etc.
You should include these queries in your list if you either sell or provide specific types of services yourself and describe them well, or if you review specific models of goods or services.
Search for a commercial offer
The last step is to search for a specific store or organization that can sell him the selected product or service. This includes natural questions like: “model haircut in Novosibirsk”, “Buy Canon 1D”, etc. They will suit you, if you have a website for a store, hairdresser, service station or something similar, then these requests are your option. If not, then they should not be included in the list.
Here, in fact, are all types of search queries. Please note: promotion for non-target requests will not bring any benefit! This is an axiom in modern times. 4 years ago this was not the case, but in modern times: believe me, search engines are already developed enough to determine whether or not your site has an answer to a user’s question. I’ll write more about search engine algorithms.
Selling inquiries
After compiling the list, you need to filter it a little and weed out the unnecessary. In particular, it is necessary to evaluate the selling power of requests, i.e. the ability to attract exactly the visitors you need to your website. If you provide website creation services, you don't want visitors looking for website creation tutorials or asking questions like "how to create a website." You are interested in those who write something like “inexpensive website creation” or “high-quality website creation Novosibirsk” in the search engine. It is these key phrases that will bring targeted visitors to you, who are more likely to buy something from you, order something, or bring other benefits. The situation is similar for blogs on various topics: you need those who do not want to buy a product from a given topic, but those who want to read about it, learn how to do something on their own, and so on. Based on this quality, filter the resulting list of requests.
The second important factor in choosing selling queries is choosing the most frequent ones. Professionals never take requests just out of their heads. To assess their degree of effectiveness, it is convenient to use various statistics services such as Yandex.Wordstat: http://wordstat.yandex.ru/ A very convenient statistics tool for our task. An example of its work is shown in the figure below.
Using it, you can see how many users have typed the query you entered in a month, after which you can select the wording that is entered most often, and also see something that would not have occurred to you. This will significantly improve the quality of your final list. I wrote more about the capabilities of the Yandex statistics service in the following article entitled “How to create a semantic core of a site.”
So, dear readers, now you understand that compiling a list of search queries is very difficult work and requires a fair amount of attention to detail. You should treat it responsibly, because it is the basis for all your future work, on which it often depends more than on everything else.
Best regards, A.S.
Selling additives are words that distinguish among potential buyers a group of people ready to buy a product or service. So, why are they needed anyway? As practice shows, of all the collected keywords, targeted applications bring only a small percentage. And what to do in a situation with a limited budget? Supplement sales people come to the rescue. They show the buyer's maximum interest in your product or service.
For example, if a person enters the request “Characteristics of an instantaneous water heater,” then most likely he is not ready to purchase the product. But if instead of “characteristics” we substitute any selling additive, then the meaning of the search query will completely change. The phrase “Buy an instantaneous water heater” indicates the buyer’s readiness to make a purchase.
It is important to understand your target audience. The list of selling additives that you can use depends on it. Selling additives can significantly increase the click-through rate of your ads. However, if used incorrectly, lovers of “freebies” will come to you along with target buyers.
Let's say you sell luxury wines. And when writing an ad, you use such selling additives as “Cheap” and “Budget”. Of course they are, but most users will not be interested in purchasing your product. Most likely, such ads will simply drain the advertising campaign budget and will not bring in applications.
So, why are they needed anyway? Let's imagine a situation where we have a very limited budget. And there is simply no money for testing a large number of words. What to do in this case? After all, even if you collect a large semantic core from it, not all words will work.
List of selling supplements
1. Words showing the buyer's readiness to make a purchase. These words have maximum priority. The buyer has already made his choice and most likely wants to purchase the product. Statistics show that these words will have the highest conversion rate.
Let's look at an example of how an ad with the request “buy flowers” is composed.
Example:
Selling additive "Buy"
It is worth paying attention to the fact that the search query is also present in the displayed link.
Below is a list of words that indicates the buyer’s readiness to take the target action.
- Buy
- I will buy
- Purchase
- Purchase
- Sale
- Sell
- Selling
- Order
- Order
- Buy
- Pay
- Payment
- Rent
- Rent
- Delivery
- Deliver
- Rental
- On credit
- Booking
- Design
- Hire
2. By price category. Among the selling additives, one can distinguish a group of words that indicate people who have decided on the choice of product. Most likely they are looking for the best place to buy. Moreover, the decisive factor may be not only the price, but also the proximity of the store.
Selling additive "Buy TV"
An example of an ad for the search query “Buy a TV.”
Below are words that can be used:
- Price
- What is the price
- Expensive
- Inexpensive
- Cheap
- Premium
- Sale
- Discount
- Price
- Rate
- Wholesale
- In stock
- Budget
3. By the name of the city or region.
There are niches in which adding the name of a city or region is a selling addition. For example, if we compare the queries “Caesar salad” and “Caesar salad in Dolgoprudny”, then in the first option they are most likely interested in the recipe, and in the second – delivery. Consider an ad for the search query “Internet in Degunino”. The user is most likely looking for an Internet provider that provides its services in this territory.
4. You can also highlight a number of selling additives that indicate an urgent need for goods or services. When a potential buyer wants to purchase a product or service “Here and Now”. Such sales additives are often used in delivery-related topics.
Example ad:
Selling additive "need"
List of words:
- Fast
- In an hour
- Now
- Urgently
- Today
5. A selling addition can be the addition of a place of sale or provision of services. For example, a potential buyer is interested in buying pens. Let's look at an example ad:
An example of a selling additive "shop"
From the request “pen store” it is clear that a person is interested in purchasing and is looking for a place to purchase the product. Below is a list of words you can use:
- Salon
- Online store
- Agency
- Shop
- Hotel
6. Providing a free first step can be a strong incentive to purchase a product. This selling addition can be implemented using a calculator or consultation. The main point is to interest the user in using a service or product from you, and not from your competitors.
Let's look at an example ad:
Selling additive "first step"
Below are words that can help attract buyers:
- Calculator
- Test
- Consultation
- Metering
- Miscalculation
- Calculation
- Grade
- Sample
- Trial
7. Depending on the topic, there are keywords that are suitable only in a certain direction. For example, you are providing services for transporting people. In this case, the keyword “Taxi” will be a selling additive. Let's consider advertisements related to tourism.
As we can see, in almost all other topics the phrases “Tour and Bali” would be “ ”. However, in a specific case, they will bring target users to the site.
Below is a list of words that can be used to display a specific topic.
- Ticket
- Taxi
- Contribution
- Catch up
This completes the list of selling additives. It is important to understand that no one knows the nuances of your business better than you. Therefore, try to select selling supplements that are right for you.
Otherwise it is called the semantic core
This is where you actually need to start optimizing your site. A correctly composed semantic core will allow you to save a lot of money and speed up website promotion, but an incorrect one can completely ruin the site owner and not lead to visible success.
What is the semantic core
This is actually a list of search queries for which we will promote our site. Ideally, we should take first place in all of them, but in practice this never happens, even the leader of the top - pedivicia - is sometimes trampled by other resources.
How to create a semantic core
To do this, in RuNet, in the general case, statistics on Yandex search queries (Wordstat) are used, since Google doesn’t have it anyway (there are AdWords statistics - but this is a little different). But this is not so scary, because there are lies, there are big lies, and there are Yandex statistics, which, of course, lie as they say. There are a lot of manuals for creating a semantic core on the Internet, and we will just briefly outline the methods. Let’s take, for simplicity, a certain site that sells brooms.
First, we need to define the selling queries. You can, of course, ask your competitors which queries generate the greatest conversions, but we are afraid they will send you to a site of a symbolic direction.
Therefore, we will leave this idea, although there is an option to dig into the open statistics of competitors, which are sometimes still available. You can simply study meta tags from successful competitors; it is clear that they will write convertible keys in the headings; some still naively list them in the keyword.
But, Shurik, this is not our method! And we will work spherically in a vacuum. Without regard to competitors. Moreover, it often happens that they are idiots and promote sites using an incorrectly selected semantic core.
To begin with, we will enter our basic word “broom” into Vodstat. And we’ll get a bunch of options, with what and how many people were looking for it per month. Now let’s turn on our brains and select from them those that our potential buyer is most likely to enter. It is unlikely that our buyer will enter into the search bar “download broom” “broom without SMS” “broom video” “naked girls whip themselves with a broom in a bathhouse”
This means that we most likely still have requests from +buy, +price +Moscow (or +Uryupinsk, if we trade in Uryupinsk). We will work with them.
Types of search queries
This deserves a separate article, so let’s briefly say that search queries are divided into:
- HF - high frequency
- MF - mid-frequency
- LF - low frequency
The division is conditional - because in one subject the RF can have a frequency of 1000, and in another - 1,000,000.
To simplify, we can assume that HF is basically a one-word query, MF is a two-three word query, and LF is three words or more (usually up to five). There is also such a term - Super LF - they are typed in the search once a week and they consist usually five or more words.
When counting words, stop words are not taken into account, usually prepositions, that is, “where to buy a broom” is a two-word word, and “buy a mop inexpensively” is a three-word word.
Main mistakes when compiling the kernel
Selecting non-selling requests
The main mistake is choosing and promoting non-targeted and non-selling requests.
The site owner and African optimizer, as a rule, as a result of formal logic, believe that since the site sells brooms, then it is optimal to bring it to the top for this particular request.
Of course, there are buyers too. But there are not many of them. Optimizers are often convinced of this and, through incredible efforts, bring “broom” to the top, and then an epic fail occurs - there are only one or two conversions to the site for this request. (And not only for the above reasons.)
At the dawn of optimization, when search engines were still stupid, the entire top list for high-frequency queries, however, was clogged with commercial sites (it is still clogged with individual monsters like “air conditioners”).
This made the user incredibly upset. He is looking, perhaps, for an article about brooms in order to write an essay on it for school, and they offer him to buy this “broom”.
Therefore, search engines, concerned about selling advertising with sane results for such queries, began to show information sites in searches, which made the optimizer’s task even more difficult.
In summary, we have - bringing a high-frequency query to the top is expensive, all sorts of pedivics and dictionaries get tangled underfoot, and there are few clients from it. Therefore, it is simply unprofitable, and not worth the effort and money to engage in such masochism.
Misuse of Statistics
When we ask Wordstat for our query “broom” and see that 99,999 people asked for it, this does not mean at all that it was the words “broom” that such a number of people were looking for.
This figure shows how many phrases using our broom were entered by users, in any case, conjugation, tense and number. Therefore, you should use refinement search operators. For example, the operator
It will exclude all other words (how to sweep, buy..), and such a construction
It will clarify the word form, that is, it is a broom, and not a broom or a broom.
Thus, it is necessary to use clarifications for the entire semantic core in order to avoid bringing to the top queries that seem to be asked by a lot of people - but in fact there are negligibly few of them, or none at all.
Selecting overlapping search queries
Everything is simple here. This is not even an error, but makes progress very difficult. It is no secret that many words in the Russian language have several meanings: for example, a faucet is both a water tap in the kitchen and a crane, a sink is a sink in the kitchen, a car wash, and the very process of washing dirt off anything. There are many examples, there is no point in listing them. The task is also complicated by the names of literary works, films, as well as all sorts of brands.
If you named your company Everest, be prepared for the fact that in the top, in addition to information about the mountain, there will also be other Everest LLCs and it will be difficult for you to advance by the name of your company. And this is sometimes very important. Your entire advertising campaign on TV (for a lot of money) is “Ararat travel agency - the best tours!” will crash in the delivery, like old man Noah’s ark on Mount Ararat, and if it doesn’t crash on the mountain, it will drown in the port wine of the same name.
What to do about it? - avoid or accept. I know real cases where companies on the Internet were forced to change their name because of such “Ararat”