This question reminds me of a blog I saw before, titled "100% guaranteed loss account to improve ROI by 50% in two weeks." The author's gist is: If your SEM account consumes 500 yuan in two weeks and generates an income of 100 yuan, it means a loss of 400 yuan. Want to cut your losses quickly? No problem, stop your account and send me 500 yuan. I promise to return you 300 yuan within two weeks. In this way, you only lose 200 yuan, a 50% improvement! The blog post uses an ironic tone, but the problem it points out is very practical: there are many customers who continue to invest in SEM even though they know they are losing money, and at the same time they suffer from the inability to achieve the ROI indicator, which is contradictory in itself.
"Keywords are not converting" is a common problem. I will list a few verification steps I may take when encountering this kind of problem. I believe Hewei has already done most of them, but they may still be meaningful to other readers. .
1. Check the tracking methods. If there is a situation of "no conversion at all" that goes against normal cognition during SEM advertising, the first thing to do is to eliminate technical obstacles. The effectiveness evaluation of online marketing is based on effective tracking. Or writing cookies through JS, or reading references in website logs, all are achieved through the process tracking of display-click-results. In Hewei's case, the brand word has converted, but other words have not. You need to confirm whether the tracking deployment of the brand word and other keywords is consistent. If there is no problem with the tracking method, proceed to the next step.
2. Check the landing page. Most customers will not use different tracking methods for brand words and other keywords, but it is very common for landing pages to be different. Brand words generally point to the homepage of the website, while other keywords may point to related pages. So is there a problem loading the relevant pages? Are there any issues with getting from the related page to the purchase page? In practice, I have not encountered common problems with non-brand keyword landing pages, but it is not uncommon for some keyword landing pages and next-level pages to have problems. This is often caused by adjusting the website structure and URL format. . Thousands of pages in large websites often involve different departments, and adjustments to certain departments or functions may have a chain reaction. If the website structure is unscientific and the URL format is not uniform, adjustments may be prone to errors, causing some pages to fail or subordinate links to fail. If the check finds that the landing page is OK, continue to check the conversion page.
3. The conversion page is actually a link that is prone to problems. One possibility is that there is a problem with the conversion payment process itself and the transaction cannot be completed; another possibility is that there is no problem with the conversion payment itself, but there is a problem loading the payment completion page, causing JS to be unable to effectively write the cookie. It is also easy to distinguish between the two, because the former can be directly reflected in internal income changes, while the latter is just a failure in monitoring, and there is a gap between real income and reported income. In Hewei's question, brand words are converted but other words are not. We usually think that if there is a problem with conversion, there should be a problem with all keywords - the premise here is that the entire website has a unified conversion page. However, many websites will use different conversion pages to record different KPIs, such as different brands/registration/download/installation/purchase/return/recommendation/additional order, etc., which may have independent conversion pages. If third-party payment is used, there may be problems with jumping between different domain names. So it depends on the specific situation to confirm whether there is a problem with the conversion page.
4. All the above three are excluded, which means that it has been checked at a technical level and it is true that non-brand keywords have not produced conversions. Then compare the converted brand words with non-brand words, compare display, and compare clicks. Generally speaking, the conversion rate of brand words must be much higher than that of non-brand words, because customers have stronger purchasing motivation when searching for brand words and skip a long period of search behavior in the purchase funnel. This is why brand words and non-brand words should be managed separately in SEM management, and the ROI standards for brand words should be different. (I have discussed brand word management before, please see here.) Assume that brand words have 100 impressions, 50 clicks, and 10 conversions; non-brand words have 100 impressions, 10 clicks, and no conversions. This is Probably reasonable. Brand words have a conversion rate of 20%. Non-brand words may have a conversion rate of 5% or even lower. It takes more than 20 clicks to convert... If the overall display click rate of non-brand words is low, there is no need to worry about why for the time being. If there is no conversion, various means should be used to increase the display and clicks of non-brand words and accumulate more data.
5. Before spending more money to buy impressions and clicks, you need to do a "sanity check". That is, whether the placement of non-brand keywords is at least reasonable. If the customer is selling QQ coins, but the keyword is "washing machine", or the advertisement is "Come and participate in the iPad lottery, click and win prizes", it will obviously not be helpful in bringing about conversions. Don’t think this is a fantasy. The examples may be slightly exaggerated, but similar situations really exist. The first batch of keywords and advertisements that many clients put in are just brainstorming ideas. We can’t completely rule out the possibility of conversions from low-relevance keywords, but they are not suitable for initial placement. When you first start SEM advertising, you need to focus on the core keywords and the most relevant ads. These keywords and advertisements can be obtained by analyzing the customer's website and products, and you can also steal some of the most relevant ones from competitors. If you see “crazy” enough keywords or ads among non-branded keywords that aren’t converting, you should remove them from your account to avoid wastage. Only keywords that we expect will bring returns are worthy of our increased investment or more impressions and clicks.
6. If we see conversions on 20 or even 50 clicks, then the remaining question is how to obtain a reasonable conversion cost. But what if you still can’t see it? In SEM management we often have to face the unknown. Obviously, with existing data, there is a basis for optimization decisions. But when promoting a new product, a new advertising slogan, or a new batch of keywords, how much cost should be invested? Before obtaining the first conversion, all investments are losses, learning costs, and sunk costs. We must prepare for losses. The question is how much. Suppose a conversion is 100 yuan, should we invest 100 yuan, 200 yuan, 500 yuan, or 1,000 yuan to obtain initial data accumulation? Unfortunately, there is no set rule, and how much you plan to invest in learning depends entirely on your business decisions. No matter what the specific amount is, a tolerable cost limit must be set. Once the learning cost of non-brand words exceeds the upper limit, stop it. If you choose relevant keywords, the possibility of not converting at all is low, but the possibility of not making any money at all is very high. In the article "Lao Zhang's Moving Company", I discussed that it may not be cost-effective for small individual moving companies to launch Baidu Fengchao. Whether it is brand keywords or non-brand keywords, whether there is conversion or no conversion, as long as a well-planned and rigorous SEM placement test proves that it is not profitable, then stop playing.
In a word, if SEM placement cannot produce benefits, it should not continue. Business activities are not charitable activities. Spending money is to make money. Never do anything that cannot make money. Before drawing this conclusion, it is necessary to confirm the reason why it cannot produce benefits, because the implicit premise of this conclusion is that "the implemented SEM placement is correct." I think this is also the focus of Hewei's question.
Article source: http://www.adsem.org/298
-