This week we talked about marketing automation a the Xoogler Growth Dinner. Here are some notes around best practices and a few of our favorite tools and resources.

Automation is about plugging machines into your marketing workflow to scale up and get more done faster. Marketing automation tools can address things like:

  • Data management. Collecting, storing and routing data across tools.
  • Taking action. Using rules and schedulers to automate actions, like social posting, drip campaigns, ad placement, bid management, etc.
  • Identifying insight. Using the pattern recognition capabilities of Artificial Intelligence to make data more actionable.

Building automation into your marketing workflow

The tools selected fit into an overall marketing workflow which starts with setting goals and finishes with measuring the outcome of those goals. Automation tools often cross over, providing value at different stages. Here’s an overview of marketing workflow with a few ideas on how you can leverage automation at each stage.

Set Goals

Before you do anything as far as setting up automation, make sure you have a sense of what matters to your business as far as goals and metrics. This stage is still best handled by actual people, not machines!

Identify your North Star metrics and make all activities accountable to these. Segment has an excellent article on which metrics to use based on your growth model: User Generated SEO Content, Viral Invite Loop, Paid Acquisition or Two-sided Marketplaces. In the article they outline these steps and provide spreadsheet examples with some of the math figured out:

  1. Identify your growth model
  2. Create a mathematical model in a spreadsheet with assumptions
  3. Reduce key metrics by conducting sensitivity analysis on your spreadsheet

Campaign Planning

Campaign planning is a necessary part of identifying how the different tools will interact and how the data will be captured at each stage of the campaign.

Best practice: Create a campaign schema that summarizes the stages of the campaign and calls out what metrics will be captured at each point. The schema should define the audience, messages, touchpoints, timing, sequencing and data flow.

Examples of tools that let you create a visual schema: Appboy, Autopilot

Best practice: During campaign planning, define what the final report will look like. This is a good way to ensure that the metrics being captured will result in a meaningful report at the end. This report should reflect the driving metrics of the campaign as well as how it performed relative to the North Star metrics.

Creative Development

Campaign creative delivers the message in the formats available at each touchpoint. This can be anything from ad copy to banners, email design and landing pages.

Brand guidelines and templates drive scale here for production-level creative. Large volumes of creative can be tested using optimization and personalization platforms like Optimizely and Sizmek for banner creative.

Deployment

This is where it all comes together and the campaign plan is set up end to end. Many automation tools focus on features that help scale setup and ongoing management of campaigns across channels and throughout the user journey. This is also where tracking is set up so we can measure the metrics that matter most.

QA is an important step of deployment. Once everything is set up, it’s important to test not only the expected experience, like whether the landing page looks right,  but also whether the data is being captured correctly and moving across the toolset.

Measurement

As the data starts coming in we monitor to understand performance and detect breakages. There are so many unplanned events that can surface so it’s a good practice to keep an eye on things so you can catch problems quickly. In a lot of cases monitoring has to be done manually. In paid search it can be automated to some extend through scripting. You can also use Revere to automatically watch the data and send alerts about significant changes and events detected.

Finally, reporting comes into play to summarize campaign outcomes and insights that can be used for the next round of campaign planning.

Best Practices for more effective automation.

Here are a few overarching tips and best practices to keep in mind:

Reliable data.

  • Make sure you’re collecting data in a way that allows you to measure outcomes and not automating based on wrong information or data gaps.
  • Best practice: It helps to start with an example of the report you want to run. This helps you focus on collecting the right data points.
  • Best practice: Be thoughtful and consistent with naming conventions from the start. Otherwise it’s going to be difficult to understand what the data points mean and can cause extra reporting work just to synthesize these.
  • Resource: Segment’s ‘Naming Conventions: One Step Towards Clean Data’ 
  • Best practice: QA your data collection as part of your campaign/marketing QA. Make sure you’re capturing the data as expected.

Communicating.

Your team communication has to keep up with the pace of your automation. A classic example is one person running a test on a web page that impacts the results of someone sending traffic via an ad campaign, but the advertiser doesn’t know when the experiment has started or how much of the traffic is impacted. Everyone should be aware of who changed what and where. They’ll want to get a sense of how it’s impacting other areas of the funnel. Give the team timely context so they can coordinate.

Content.

  • Automation allows you to scale, but can your creative keep up?
  • Everything will move faster when have good brand guidelines and templates in place.
  • For efficiency: separate strategic creative versus production into different roles and processes. Production-level creative should be fast and relatively cheap and at some level you can automate the build out of this with the right templates in place.

QA.

Make sure you test the end-to-end flow from the user’s perspective and the data collection at each step. Data gaps or incorrect data can impact the automation and create a host of reporting problems.

More Resources!