Improving Performance with Information and Insight
Published by Lehman Associates

Three Stages of Automation

As noted in our earlier post, our adoption of technology to date has been largely focused on improvement of that which we do today to gain efficiencies and productivity. Looking forward, organizations will leverage technology for transformation to go beyond what they do today to engage with customers, members and donors. See: Automation – Going beyond efficiency

Automation plays a big role in this shift in thinking and spans across the shift from improvement to transformation. We see three stages of automation. These are not merely making increasing use of a singular technology. The technology changes, but perhaps more importantly, the mindset of the organization changes to embrace data-driven approaches to how they interact with members and donors.

Automation stages

What are these three stages and how do they differ? One way to answer that is to compare the elements of each stage – what the next action is based upon, the source of that data, and analytics that come into play to use that data to determine the next action.

Fixed is the automation that is most familiar and involves technology to support an established and largely unchanging process. It may be simple such as templates and data population to blast email a monthly newsletter or involve complex segmentation and dynamic content to send a customized version of the newsletter to various groups of member. Each member within each group receives the same newsletter and while the content may vary, the process steps are the same for all segments. Most commonly, the fixed automation process is based on existing data such as that contained within a member or donor record.

Dynamic refers to automation where the process is governed by behavior by the end user or recipient. That behavior could be a response to an email, pattern of use of a website, social media / private community postings, and so forth. Website personalization based on past use and / or purchases would be one example. One of the more common uses of term automation today applies to marketing automation. For example, a first marketing or fundraising email sent to members or donors would be the same (perhaps segmented by interest as noted in prior paragraph). However, the second email to each respondent would be based on their response to the first. Non-respondents might receive a second standard email. However those opening, clicking a link (or a particular link), or taking action on a landing page would each generate a different follow-up email for that respondent appropriate to the response. Someone who opened the email could be sent a second email that encouraged them to take a link to find out more. Those who already took that action in the first email might receive a second email that encouraged them to take a link to register for the conference or read comments from past attendees. As a result, the campaign is much more personalized. Marketing automation requires an investment of time to develop the various emails and create the rules to govern the process, but can pay great dividends in terms marketing efficiency for the sender and a more relevant series of content for the recipient. While fixed automation almost always focuses on improvement of existing processes, the more advanced dynamic automation approaches can be transformative as they enable organizations move beyond what they do today.

Predictive automation is the emerging model, anticipating, not merely responding to, needs. It taps into data beyond the member record and activity, and employs advanced modeling and analytics. A simple prediction example, used today in e-commerce, matches the user’s purchase history with similar patterns among other users to suggest products and services that might be of interest. It has been around since the late 90s. This is a fairly simple process of pattern matching, i.e., people who bought this product also bought this other product. The rapid development of artificial intelligence (AI) holds the promise to greatly expand predictive applications. Google chief executive, Sundar Pichai, speaks of an AI-first versus mobile-first as the future of the web and e-commerce. For example, many professional societies offer professional development, and frequently target market particular courses of study to members based on their career stage (such as entry, mid-career, late-career stages) or other data drawn from the member record. A predictive model might tap into local census data and economic trends to predict the future mix of services that professional may need to provide, and then suggest professional development resources to help that individual prepare. Predictive automation is almost always transformative as the technology and methodology themselves become part of the value proposition for members and donors.