As Socrates almost said: an unmeasured campaign isn’t worth launching. If you don’t have concrete goals and KPIs for your campaign (relating to profitability), then it’s certainly going to be a dead loss.
Not all metrics are created equal. While clicks and sessions might bring a smile to your face, they don’t make you money. They’re vanity metrics. Instead, it’s crucial to capture the metrics that will enable you to drive engagement, optimize your campaigns and scale.
In this article, I’ll outline three powerful, underreported metrics that will do just that – and that will allow you to genuinely enhance the performance of your PPC campaigns.
If you’re spending money to make money, then it’s likely that you’re tracking your campaign performance with basic metrics like clicks, spend, CTR, conversions and profitability.
You’re probably also following strategies such as adjusting your bids, creating negative keywords and separating campaigns between different mediums like search and display.
If you’re not doing any of this, then you should re-evaluate your ad strategy. By skipping these best practices, you’re more likely to generate a loss and be pouring money down the toilet.
While these metrics are important, revenue per session (RPS) should be your driving decision maker. You’re most likely calculating your bid price as the lowest value that it can be while still getting traction on the respective network.
However, you can only calculate profitability and relative profit margin if you know the true revenue earned per session. While setting a cash value on clicks and conversions is key, it’s more of an estimate than an exact number.
If you base your bid on the assumption that 60 percent of clicks result in a revenue-generating action (say, bidding on an auction on eBay), or that 10 percent of signups choose the paid option in a freemium model, then the profitability of your bid will be determined by multiple fluctuating, independent variables.
Spotify’s PPC landing page, for example, encourages both freemium and premium sign ups. This way, they can attribute not only user acquisition, but RPS for each campaign:
If these values prove to be radically different for a specific campaign, then they may blow your estimates out of the water. Using RPS, on the other hand, means that you’re comparing spend against revenue, like-for-like.
RPS also allows you to account for multiple revenue-generating events. For example, a conversion might count a user buying a single item, but RPS would allow you to account for multiple purchases or a purchase and an upsell. This then allows you to segment the most valuable customer acquisitions in order to maximize profitability. Be like Amazon: know where your most valuable users are coming from.
It also allows you to tie negative revenue events, like refunds and returns, to your campaigns. This way so you can establish how they affect revenue and profitability.
The key to calculating RPS is to move from binary counting to decimal – don’t count conversions as a yes or a no, but count events based on their actual cash value. Then use cross channel analytics to tie all the conversion events (positive and negative) to the session.
Another crucial metric in a PPC campaign is the rate at which users who arrive on your landing page immediately leave without taking action.
Depending on the structure of your conversion point, there are a number of reasons why users drop off. They might bounce before clicking through to a second page; fail to fill in a form; or give up on the purchase process before completing their order.
If you have a multi-step process (e.g. requiring users to click through to a second page before filling in a form), then knowing where they exit is of particular importance.
If you have a single-step process then you can track bounces against UTMs (like source and campaign) in Google Analytics and immediately see where bounces are coming from. With more complex processes, you might need more complex product analytics.
Even if you’re running affiliate programs on a CPA rate and are only paying for conversions, bounces are still worth tracking. While a poor process may not be losing you money, it won’t be making you the profits that it should be, either, and will prevent your campaigns from gaining traction – meaning that they’re a waste of your time.
There are a wide range of reasons why users may not convert on your campaigns:
Split testing CTAs, page design and the copy on the page is highly effective in identifying the best page layouts to drive conversions. For example, Marissa Mayer famously split tested 40 shades of blue to get just the right color for hyperlinked text. Uber’s current AdWords landing page for new drivers has also been tested extensively, from the right form fields to the CTA:
While this method of optimization can be practical, sometimes the results are marginal. Therefore, you may want to consider a complete redesign of your landing page, taking the entire journey into consideration.
While analyzing why the winning variation performed better, it’s important to understand the behavior of your users. What is it that compels your users to act? Is it the value proposition, or perhaps the UX of your new landing page? This understanding will lead to breakthroughs in other areas of your marketing.
If you want users to perform a specific action, it’s sensible to strip back distractions and alternative options and to point them very directly at a single CTA. For an example of clarity of purpose, look at Google’s homepage (below) and then compare it to what’s on offer at Yahoo.com. It’s also crucial to match your messaging to the interests and expectations of your audience. Speak to them in their language.
Products like Optimizely, Hotjar and Convert.com make A/B testing easy, and include tools to make edits to your pages on the fly without the need of a front-end designer. Meanwhile, heatmapping tools like CrazyEgg (indicating where users are clicking on your pages), can illustrate how your audience is engaging with your site and how the design can be tweaked to push them where you want them.
Bounce rate can also flag issues with your campaign. For example, if errors in your campaign setup mean that you’re targeting a UK audience with a French-language website or directing mobile users to a non-responsive page, then they’re unlikely to be able to work out, or care, what you want them to do.
In some cases, however, bouncing users may not be a complete loss. For example, if users have followed several steps of a purchase or conversion process, then there may be significant value in retargeting them.
Networks such as AdRoll allow you to integrate with CRMs, which is a nice way of adjusting your messaging based on lifecycle stages and other attributes.
Of course, you may have to balance multiple factors to get the best outcome for your process. For example, when directing a user to purchase a product you might find that a significant percentage of them bounce at a late stage when they see the shipping pricing and options.
While you may want to solve this, it should be weighed against the drop-off rate if P&P details are presented upfront, at the beginning of the process – which may actually be much higher. Split testing the two user journeys can provide you with the definitive answer.
The more you know about your users, and the users who convert, the better you can target your campaigns. This can include all of the following:
Targeting different audiences is a basic best practice, as is excluding users who are ineligible for your offering. For example, if you only ship nationally, don’t advertise abroad.
It’s wise to challenge your assumptions. Your health supplements may appeal to bodybuilders as well as retirees, for example. This is something that you would never know without testing your targeting options.
The takeaway from this is that you should create specific campaigns and ad groups for different audience segments. While it may well be tempting to compare stats between audiences, it makes more sense to treat them as being fundamentally different. Develop unique campaigns, landing pages and even test different pricing structures for each subsection of the market.
You can effectively pull user information from either end of the process. If you select your audience at the campaign level and target specified landing pages, you can be confident knowing who is doing what.
Equally, you can pull significant amounts of browser information once the user is on the page and you can extend this with purchase information (and information gained from forms) – before ultimately tying everything together with tools like MixPanel.
Coming from the other direction, as you gain more information about converting users in each segment, you can tighten targeting, in order to increase CTRs and conversion rates and reduce campaign costs.
You’ll also get insights across different markets about the level of competition in each. You may also discover that a peripheral audience is actually far more profitable for you than what you had presumed was your core market.
If you have a tight grip on these three values then you’ll have an extremely detailed understanding of how your campaigns are working and why – and of what you can do to enhance performance.
If you’re buying at scale then you may also want to build in automation to cut downtime. This will also ensure rapid responsiveness and simplify processes where complex calculations are involved. Track, optimize and iterate – and head onwards and upwards.
Juuso Lyytikkä is the Head of Growth at Funnel.io. Funnel is a marketing analytics tool for online marketers that collects data from ALL advertising platforms and allows marketers to send and visualize this data anywhere.
Please read our Comment Policy before commenting.