Data Driven PPC: The Secret To Paid Search Success is Quality Data

Quality vs. quantity. Nowhere are the tradeoffs more important than in the world of search engine marketing. Specifically, with the data that we use to measure performance, set goals, and drive the decision-making for our paid search programs. As a fellow SEM professional, you know that data is everything. It measures our value. It allows us to define strategies and boost our performance. It is the driving force behind the work that we do. But data on its own is not -- in and of itself -- an automatic key to higher value, better performance, and data-driven PPC. The operative piece is actually quality data.

Consider this: an athlete is dependent on quality food to help create and maintain energy, increase endurance and build strength. If that athlete indulges too long or too much in foods that are high in sugar, contain simple carbohydrates and other additives, their performance will inevitably suffer. Chances are they’re going to be slower and weaker -- and even micro-seconds of lost time will inevitably benefit their competition.

The same applies to how key decisions are made for running a large-scale SEM program, of which no decision is more important than defining and executing on a bidding strategy. Quality data is the key requirement to enabling advanced optimization strategies, fueling accurate forecasts, and activating key insights -- insights that allow you to have visibility into your sales funnel and the entirety of the customer journey. In turn, this not only enables you to provide a more relevant and valuable customer experience, but allows you to drive the optimization techniques that can boost conversions, gain an edge over competitors, and ultimately increase revenue and ROI.

Perhaps not surprisingly, missed performance goals often boil down to incomplete or low-fidelity data that drive misguided decisions. Simply put, if you don’t have quality data, it doesn’t matter how sophisticated the algorithms or optimization techniques you try to apply to your program, your performance will fall short of its peak potential. To put this in perspective, the lack of a complete and unified quality data set represents a significant challenge for most SEM programs.

However, solutions that can capture and leverage the right quality data can have a very meaningful impact on the success of an SEM program. Here are a few ways:

Driving Performance that Leads to Higher Revenue and ROI

Now more than ever, data has the ability to directly drive increased business performance. By measuring, tracking, and analyzing data on past performance, smarter and more effective decisions can be made to deliver improved results moving forward.

In fact, certain core categories of data that most businesses already have can be directly applied to increase SEM ROI and drive higher revenue. By tying analytics data from web and mobile tracking solutions, offline data from call centers or CRM platforms, inventory and/or capacity systems that track real-time supply constraints, and contextual data like micro-weather signals that influence conversion rates back to SEM data, search teams can empower smarter and more profitable bidding decisions. Few, if any, large scale SEM programs can achieve peak performance without leveraging data from these types of sources.

In addition, a unified data set will not only offer insights into how deep your prospects are in the funnel, but also illustrate the various milestones of the customer journey to which value can be attributed. If you don’t typically have conversions for at least 30 days after an initial click, for example, a unified data set can provide insight into other stages along the path that can serve to predict not just whether a click is likely to return value, but how much.

Conversely, if your SEM program isn’t achieving peak performance, this likely means you should evaluate your SEM data mechanics. At the highest level, you should ask if your current bidding solution is capturing and leveraging all the critical data that measures and influences your funnel, regardless of where that data resides. For example, if your solution is having trouble integrating data from third-party sources, offline locations, and/or deeper funnel metrics -- or simply can’t do this -- you may be forced to use a sub-optimal strategy of optimizing bids to metrics that don’t actually measure business performance.

Here are some other questions worth asking to evaluate your data mechanics:

• If you’re using a newly deployed bid management platform, does it integrate and apply enough historical data to fuel intelligent decision-making right away?
  
  
• If your click to conversion funnel is days, weeks, or longer, does your bidding solution intelligently address the lag and make informed and proactive decisions based on predicted performance?
  
  
• Is every reported metric eligible to be a performance goal? Often times solutions may present a relevant business metric in a report but not make that same critical metric available as an optimization goal.

Achieving peak performance is hard. But more than anything, the foundation to achieve it requires investing in a platform that collects the right quality data and then allows you to unlock its potential by applying sophisticated algorithms to automate bidding optimization at scale. Being data-driven in paid search demands it.

Building & Leveraging Historical Context for Better Decision Making

And what if you’ve just deployed a new bidding platform? If you want to minimize the all-too-common and expensive “bidding learning period,” the quality of data you feed into the new platform is of paramount importance. At a more tactical level, if you don’t integrate historical data, you and your team may struggle to see patterns, garner key insights and make the most profitable strategic decisions until you build a new history. Here you would have a much more manual process of finding out, for example, whether May was a stronger month than June, and for what reasons. So when it comes time to set budgets and give forecasts for next Q2, they will be more difficult to prepare and less informed.

Why else might historical data matter? In the world of SEM, almost every business experiences some sort of seasonality, whether it’s based on promotional events, holiday periods, or a combination of both. Ideally, your future promotions and seasonal strategy will be driven by lessons from past successes and failures. Likewise, your bidding strategy should also be able to leverage data from past promotions and seasonal periods to intelligently guide future bid calculations and adjustments. With the right platform, strong historical data can provide the context and insight necessary to make proactive decisions to maximize business value.

Thus, it’s not only important to access and leverage your historical data, but also to be confident in the fact that the data set feeding your platform will be put to full use. That is being data-driven in your PPC. If your platform only uses 30-60 days of historical data for bidding, for example, how can it account for a once-a-year event like Black Friday? What ends up happening all too often is that paid search professionals are forced to completely override their bidding platform and control the bids themselves in order to be confident the decisions being made will best align with business objectives during such volatile periods.

While it might be a lofty goal to achieve “full automation,” today’s paid search marketers need a solution that provides robust data integrations and the capacity to fully leverage that data in a proactive way to win in a competitive ecosystem and achieve peak performance. Among other capabilities, this should include the ability to integrate and report on promotional and seasonal data from the past, along with ongoing revenue and performance metrics all in one pane of glass.

Enabling Value-Added Techniques That Differentiate You From Competition

Remember how we talked about the quality of the food that athletes consume, and how it directly impacts both their potential and performance? Quality data operates the same way in terms of powering advanced techniques and capabilities in your bidding solution. These are the features that ultimately enable you to differentiate yourself in a competitive market and can take you to the level of achieving peak performance.

Take forecasting and scenario modeling -- while the concept of predicting the future isn’t new to the industry (and is something Accounting and Finance teams never stop asking for), it’s a far more accurate science to do when fueled by reliable and comprehensive data sets. The output of algorithms that make such calculations within a bidding platform will be empowered -- or limited -- by the depth of data that the platform ingests from the sales funnel and customer journey. But make no mistake, capturing all available data points will provide a richer understanding of a click at any given point in the process and better enable modeling as to how that click’s future will transpire, even when accounting for a delay of days or weeks to mature into a deeper funnel sale.

Similarly, automating audience bid adjustments is another example of advanced bidding techniques that enable you to outperform your competition. Status quo would mean relying on humans to define segments, collate and analyze the data, and then attempt to determine the appropriate audience bid adjustments to make and when. On the other hand, by automatically tracking all the audience attributes of each click and understanding how those attributes impact conversion rate and monetization, a bidding platform can be empowered with the raw fuel to be able to drive an automated approach. And rest assured that with proper data science methodologies used to model this data, it is possible to unlock higher performance and garner the greatest value from customer audience data. That’s a key requirement to achieving peak performance.

In Summary

Humans vs. machines. It’s an age-old question. Historically, SEM teams have been responsible for not just the performance of their advertising spend, but also coming up with the tactics, processes, and tools to achieve it. Today, technology now offers the ability to go well beyond what humans can accomplish on their own or even with the assistance of first-generation bidding platforms. This means the best of all worlds: humans plus machines. Being data-driven in paid search isn't possible without this machine side.

What’s the alternative? Regardless of the size of your organization, at scale, there aren’t enough humans to review all the relevant data and execute all of the necessary decisions to manage a program with hundreds of thousands or even millions of keywords or products -- you simply can’t scale your team to accommodate all of that data. And if you try, you certainly won’t achieve peak performance. Enter technology, which can do both the blocking and tackling of keyword-level bidding, as well as more advanced strategies like device, location, ad schedule, or audience bid adjustments. But as we’ve discussed, that automation can only begin with strong, reliable quality data.

High-powered SEM platforms have the ability to provide a critical link to understanding where your business derives its value while also forging a path to increased ROI and higher revenue. They also provide historical context that enables you and your team to identify patterns and make strategic and informed decisions quarter to quarter, week to week, hour to hour, and even minute to minute. It enables you to have truly data-driven PPC programs. But, like the fuel that drives a sports car or the food that fuels a professional athlete, performance doesn’t come without the right inputs. In SEM, that means peak performance can’t be achieved without reliable and accurate quality data.