What You Need to Know for a Successful SEM Campaign

What I’ll discuss in this post should be hard-coded in every SEM-expert’s DNA, but since I found that a lot of people out there struggle in approaching Search Engine Advertising with a true holistic view, I thought it was worth spending a few words on it.

The content of this post originally comes from a slide a created for a training. When I showed it to a few colleagues to have some feedback on my “theories”, one of them said: ”Wow! This makes look what we do so complicated that if you give this slide to a media agency, they could virtually charge any fee for their service!”.

As a matter of fact trying to embrace all the different elements that make Search Advertising campaign successful in one single slide (and keep it simple at the same time), is quite a challenge! Moreover I feel like I’ve never been blessed with a superior data visualization ability that some people have, and let’s face it… my slide is pretty horrible for the time being!

I’ll let you be the judge…


So if you think you are able to represent in a nice visual way what I’m going to describe in the next few paragraphs, well my friend… we are in business!

But let me break down my thoughts in to the different steps of this process.

First a little nomenclature though:

  • The “=” should be interpreted as “is a function of
  • When a word is red means that the metric can be broken down in to its component;
  • When a word is blue it means that the metric can be directly influenced in the short term;
  • When a word is black means that it is either the metric under discussion, or that is not influenceable in the short term (i.e. overall value of the offer).

Caveat: the model takes for given exogenous components such as brand awareness, brand recognition, competition and so on, even if they are obvious influencers of the outcome of the advertising campaign (i.e. see for example when Apple came out with the iPad2, the quality of the advertising had a relatively small weight on the number of sales with respect to the Old Spice’s body wash).

Here is what you need to know for a successful SEM campaign:


 But let me break it down and explain it.



A “lead” is not always an “acquisition”

It won’t be hard to argue that acquisitions are a function of the number of conversions you are able to get from your website, minus the dropout rate (if you sell products the dropout rate could be cancelled orders, if you collect leads the dropout rate can be the ones that are not converted in to sales or customer acquisitions).

The dropout rate can be influenced by many factors, like shipping costs higher than expected, hidden fees, offer not attractive, etc. The way to find out what influences the dropout rate is through a knowledgeable use of analytics, customer surveys and benchmarking. We can influence the dropout rate in many ways, by fixing the problem(s) that is causing it, or through advertising techniques like re-targeting, direct mailing, and so on.

Let’s see now how we can influence the conversions number.



Obviously more traffic we have to the site, higher the chances of having conversions. This all depends though by the ability of our site to convince people to convert. Always remember that increasing your traffic through paid search, is only one of the two ways of increasing conversions. Increasing your conversion rate by 20% may cost you the work of your analysts, IT specialist and webmasters, but after that it’s free. And most of the times you see websites with a conversion rate so low that moving the conversion rate from 2% to 4% (100% increase in conversions), costs way less then a 20% increase in budget (20% increase in conversions / given a constant conversion rate).

I admit the cost of smart and capable people is not low, and there are many highly paid “experts” out there that have absolutely no idea of what they are talking about! But as Avinash Kaushik always says: you should spend 90% of your marketing investments in people and 10% in technologies.



As I was saying before, increasing the conversion rate of your site takes some acumen, but there are three fundamentals levers that we can action in order to achieve this goal.

  • Value of the offer: higher is the value-for-money you are proposing, higher will be the chance that people will buy your product.
  • Readiness to convert is something that we cannot influence directly, and it require us to take to account the average time-to-conversion that is required for our customers to finalize the purchasing decision. Economic theory tells us that higher is the value (monetary or emotional) of the good we want to purchase and the complexity of the product, longer will be the time to go through the purchasing funnel from awareness> consideration> intent to purchase> purchase.
  • Website usability: in every single step of the conversion funnel our website and advertising campaign can influence the user to move to the next step. In order to do so we have to have a website that is optimized to guide the user towards the final acquisition objective. This means having a website that makes all the information that the users need to shape their purchasing decision available (but only the necessary ones!); it means also making the conversion process seamless and optimized for an easy checkout (e.g. allow checkout as a non-registered user). While is harder (and longer) to influence the first two levers, anyone can start working right away to improve their website usability. The most sophisticated web-marketers will also find a smart way to monitor “micro-conversions” (all those value-added activities that the user can perform in our website), in order to be able to re-engage with the user at a later stage, both to decrease the dropout rate or to favour upsells and cross-sells.


Visits to our website are obviously made of the number of clicks on entry links (let’s just look at SEM campaigns for the time being), minus the bounce rate, which is made of the pople that decide not to engage with the website and leave right away. Once again, in order to increase our acquisition numbers we can either increase the number of people that enter our site, or decrease the number of people that leave without converting.



The bounce rate is a function of traffic relevance (see below) and landing page quality at the moment of the first impact of the user with it. This is highly related with the expectations of the user that is entering our website and are usually shaped by:

  • the promises that the ad message which brought him to the site has made (let’s consider only SEM);
  • the expectations with respect to the needs that he wishes to satisfy through the visit to the website;
  • the benchmark with respect to the standards set by the competition (no one expects to pay to get a quote for the service: specifying that a quote is free will encourage more to engage that having to pay for it);
  • the technological standards of the industry (no one would be willing to wait more then 3 seconds for the page to load, as well as no one would expect to find a Windows 95 style layout);

To increase the probability that a user will converti is also important to link a specific group of keywords with the most relevant ad and landing page: in Google AdWords this is done through a wise selection of AdGroups within your campaigns, that will show a tailored ad messaging and landing page for a specific group of keywords.



The bounce rate is influenced by the relevance of the traffic brought to the site: you can advertise on every combination of keywords known to mankind, but if thy are not relevant to your business, your are just driving up your costs and you are spamming people!

Moreover you have to be careful that your ad doesn’t promises something that your site can’t offer. For example, if I searched for a specific model of a product, I saw your ad that was talking me about that specific model at a discounted price, when I get to the site I expect to see that exact model whith the discount you promised me; otherwise I’m going to feel deceived by your ad and I’m going to leave your site (bounce back), possibly hating you with all my heart!



The ad plays a fundamental role both in the quality of the traffic I drive on the website, but also on the number of clicks I get. Even if you bid only on the most relevant keywords, you may still not attract the proper amount of traffic due to the fact that your ads are not appealing to the users. Catchy phrasing must be coupled with the promise of a valuable selling proposition that gives an edge to our product with respect to the competition (promise that must be kept at all costs and clearly visualized in the landing page). Limited discount? No shipping costs?  I want to know about it right away, but they are going to make me choose your product over a competitor’s one only if the offer is truly unique! If no one makes you pay for shipping costs then advertising it won’t give you an advantage, but it is still necessary requirement to compete in the market. [For the fanatics of economic theory here we find the concepts of Unique Selling Proposition and Ideal Point of Offer]



Obviously we don’t need to forget that the number of clicks are fundamentally dependent on the number of impressions your ad gets times the CTR (see below for its components).



Even if this may seem counter-intuitive (and wrong), the reality is that I’ve got coght up so much with the mathematical formulation of this “theorem”, that I’ve applied a substitution. If you remember a few lines above I’ve assumend that the traffic relevance was a function of keyword relevance and the attractiveness of the ad.

Having said that here I imply that the click through rate is a function of keyword relevance, attractiveness of the ad (value proposition, messaging, etc.) and ad position. Needless to explain that higher ranked ads get haigher visibility therefore have higher chances to be clicked (see the F shaped heat map that that shows the attention given to the elements on a search results page on Google). This doesn’t imply though that a top position is the one that will allow you to have the greatest ROI or profitability (quantity x price – cost of good sold).


The attention given to the search engine results page follows a F shape

While we can directly influence the bid we are put forward for a specific keyword (adgroup or campaign), quality score is a mysterious and complicated metric to influence.

One of the main elements that influence quality score is the historical CTR. Obviously if you bid on a keyword like “chipset” when you are selling designer begs, your quality score will start low. But the system of a (good) search engine is always eager to learn from it’s user: if it sees that people keep on clicking on your ad very often (keeping a decent ctr), it will gradually improve the quality score to the point that will “learn” that a famous designer has created a new pochette called exactly “chipset”!

(If we want to get very picky about the silly example I just made, let me say in my defense that “chipset” will still be a very bad keyword to choose. I would go more on keyword like “chipset bag” “chipset pochette”, “chipset -name of the designer-”, and use wisely match types and tons of negatives!)



This is very straightforward. Both levers can be actioned immidiatly. The only thing to notice is that in the Keyoword # I imply also the “nature of the keywords chosen”, which includes both how generic they are, and the matching option chosen. Obviously we need to remember that greater the number of generic keywords chosen, greater is the chance that the relevance of such keywords will decrease, compromising traffic relevance and quality score. [Let’s have a postulate for this: keyword number grows in inverse proportion with respect to keyword relevance in an accelerating fashion]




When managing a SEM campaign there are many switches that we can play with right away in order to influence the success of our online advertising campaign.




Some of these metrics, such as budget, keywords number (considering also their broadness and match type), bids, keyword relevance, messaging, landing page quality, website usability and dropout-rate, can be influenced in the short term in order to see an impact in the success of our campaign.  

[Caveat: obviously this implies that our online marketing department has full control on the website (landing pages, forms, carts, etc.); often happens that the website is managed by the IT/engineering department, which doesn’t always implement the things for the benefit of your advertising campaign. If this is the case make sure your CEO realizes the missed opportunities (profits) deriving from a sub-optimal website, so that he can grant you a better spot in the priority list of whoever manages your website!]

At the same time we have to remember that in case of faliure of our SEM campaign (negative ROI, high CPA, etc.), we cannot blame it only on a handful of metrics, but we need to consider the above function holistically with respect to all the elements that constitute it.

Moreover is important to remember that the cost/benefit ratio of an improvement to one of its components, shouldn’t be confused with your ability to implement such a change, but should always be evaluated in light of a sound analysis. This means that if changing the layout of the landing page of your campaign is going to be painful just because who’s in charge of it has other priorities, doesn’t mean that it’s not what it needs to be done, but just that your campaign lives in a suboptimal business environment and is destined to bring a suboptimal result.

Obviously I understand that in no company there is an ideal business envirenment and everyone needs to compromize between costs, resources and results; but is always fundamental to remember what is the difference between what can be done and what should be done. I believe that doing everything that can be done to achieve what should be done will make you not only a better online marketer, but a better person overall.

Author: Paolo

Economist by education, marketer by profession, coffee roaster by hobby.

5 thoughts on “What You Need to Know for a Successful SEM Campaign”

  1. Hi Paolo!

    Wouldn’t you mind if I translate this article into Russian and post this translation to my personal blog dedicated to business analysis, startups, advertising and other stuff? Of course, I am going to give a link to your original article there.

    Thanks in advance.

Comments are closed.