You might say that a SaaS company is only as healthy as its acquisition channels are. This is a story about how we handle multi-channel acquisition and an overview of everything we do to keep those channels flowing.
User acquisition is a potent part of marketing - that implies advertising, tracking user behavior during various phases of the customer journey, and acting in accordance with their expectations. It is exact, as it uses only clearly defined metrics and its goals are very tangible. In fact, customer acquisition is Real Work happening fast and live.
Responsibility is an important issue here, since - in performance-based marketing, the resources justify the budget, as long as CAC (Customer Acquisition Cost) is below the profit margin.
Data helps us to learn, to grow.
We often hear about email lists being sold, various leads which would suit us, but in the end - victory is rarely achieved by such means. When someone does manage to succeed in this, the entire world focuses their attention on them. A company that succeeds, does so because they knew what they were doing. They were more proficient at this than their competition, and understood how to properly gather, read, analyze data - and what to do with it. The actions of such visionary companies change the world every day.
There’s no room for bluffing in an exact science.
Therefore, we strive to simplify this process by looking for and using common and recurring factors.
Every business lives and dies according to their profits. The amount of budget we have at our disposal represents the channels of acquisition. Each channel has its own rules and particularities, but all of them have the same goal - to bring in new users.
The performance of the channel in the overall result is called attribution. It shows us how big of an impact an individual channel had on the result.
When done correctly, attribution can provide us with a clear and accurate insight into how, when, and where marketing affects buyers, and on which devices and channels. We use that info to spend more wisely and define the optimal blend of client interaction. In a nutshell, with channel cross-rating we can do more with less, because we understand our clients better.
The problem is that many still underestimate the difficulty of such advanced analysis on a company-wide level. The majority still use outdated methods which have become obsolete in the complex ecosystem that is online marketing. Some methods, such as “last click”, have been discredited a long time ago. To do all of this the right way - which is a holistic approach - requires not just great data, but also great criteria, and great models that can handle such problems.
Let’s look at a hypothetical user that is conducting a search for a specific keyword. The user sees that we come up first in the search results, because we, for example, have paid advertising, and visits the ActiveCollab website. After a couple of minutes, the user goes off to search some more. The user forgets they even searched for us, but our remarketing system reminds them of this later by showing them our ad on a site they use for searching.
To analyze these metrics, you’ll need to set up some tracking cookies, remarketing scripts or pixels, and to have a systematic, defined naming of every item, campaign, and source/medium you have.
For example, in Google Analytics we have created a custom channel grouping to track our global campaigns, but then we realized that such a setup can’t be exported, because export is only possible with the default channel grouping. We suggest either to change the default channel grouping, or to make a new analytics view and try to reconfigure the default channel grouping there, as well as other custom dimensions you want to track. The point is to always have a backup of your default analytics settings.
Concerning custom dimensions: advanced tracking can’t be considered advanced without them. Let’s say you have a subscription form, and those inputs are pushed to your database. In the custom dimension you can push back that particular user_id from the database, and combining it with a secondary dimension - you can find out the source or the campaign in question that brought that customer.
Another thing, a visitor can arrive at your website many times - as we’ll see next - so track those sessions. Add session_id, and an additional time_track as a custom dimension. Now you have a source/medium, campaign, user_id, session_id, and a time_stamp for every instance that you track. Now you know where your users came from.
The same user can click on the ad, but more often - they search again, only this time they reach us through an organic search, because we are at the top of all search results. Also, the user decides to read a review of our software, as well as to check what social networks have to say about us. From Facebook, they once again arrive at our website, and subscribe to our newsletter. The user likes what they see, and they visit our website directly, they show our product to their manager, who then purchases the software.
This type of behavior is becoming more and more frequent. Gone are the days when one good word was enough to base your entire decision on whether to buy something or not. Because of this, all of us who work in marketing have to assign value to every individual step of this process. We’re careful not to assign individual value through ROI per channel, but to improve the ways we distribute the budget to all the channels that are part of the buyer’s journey, even though some channels don’t have a great score - when looked at individually. Multi-channel attribution provides us with such data.
The key words here are the “big” three - big data, big measurement, big system, and tracking.
Google is tracking your behavior, it “gives us data” on how serious you are about purchasing and using our software. We then calculate this intent. If you’re interested in how we do this - we can discuss it in the comments section.
At the same time, Facebook is following your habits, your interests, and it allows us to find more of your lookalikes - people who display behavior similar to yours. The key word here being - user behavior, from the moment you first learned about us, to the moment you bought our product.
Linkedin, Twitter, Quora, all of these, with their own sets of specifics, contribute to the final result - sales.
Facebook can track user behavior from websites through Facebook Pixel and another script called “Event code”, which we use to track “add to cart”, “view content”, “purchase”, “complete registration”, etc. With this, we can create custom audiences to include in or exclude from ongoing campaigns.
For example: for most of our campaigns we’ll exclude anyone who has already been acquired by the retargeting pixel. This way primary campaigns are only reaching new potential users who haven’t yet visited ActiveCollab’s website. Then, we run an additional campaign that separately retargets all the past visitors to our website.
This entire road of every specific channel acquisition represents a multi-channel funnel, and it shows us a bigger picture that illustrates how the user makes a conversion. Of course, we have a perfect scenario of how we would prefer our visitors to behave when they visit our site, and in most cases - they don’t adhere to it. We use all sorts of tactics to lead or bring back the users to the path we want them to take.
Each of these marketing channels spends a portion of the budget and, in proportion to the participation in the result, sums up its performance for each user which was brought in through it. The sum of all of these amounts for one user represents the CAC, meaning - the cost we paid to get to one customer.
This set of data is monitored live through the campaign monitoring report, where we clearly see the effect of each marketing channel, then each campaign, and the ads of that channel.
The experiences of buyers are becoming all the more complex, the influences of marketing channels all the more intertwined. Consuming content is now fragmented, the “attract, engage, retain” model that was once made up of three steps is now divided into 37 different parts.
With all of this data, we endeavor to make every step of the journey clearer and more worthwhile. Data is the new oil, but we aren’t selling it. It’s a unique material that requires special care and attention, and gives us access to solve problems and create new opportunities.
Your turn now. Which multi-channel acquisition problem are you trying to solve?