Demystifying real-time analytics: a brief primer
Demystifying real-time analytics: a brief primer
This often misunderstood (still) novel area of analytics is - applied thoughtfully - highly actionable for analyst work & automation of business outcomes/better UX
While many organizations are still learning to be data-driven and find the right talent to staff their teams, those who are advanced continually push the envelope with what is possible. A (still) emerging area of analytics that’s frequently mentioned, equally misunderstood, but infinitely actionable is real-time.
Media company analysts are for sure some of the most fluent with real-time, with the most popular newsrooms at the largest sites we visit frequently using (real-time) data to determine which stories are worth promoting on their homepages, featured sections/widgets of apps, etc. This, in tandem with giving hot, potential break-out stories a push in social/to influencers, email readers and partners all happens in unison with goal of increasing the likelihood their narratives rise above the clutter and define the world’s (or in some cases simply an industry’s) Zeitgeist of a given day. I worked closely with teams at several household names in the space while at Google, helping them get the most out of our tools, and was always impressed with both their level of sophistication and dedication to trying new ideas here. Not just media, tech companies as well: a great example here is Twitter, which uses a hybrid of real-time automation via analytics as well as human curation to populate their trending topics section. The use cases there are clear, and what we see daily online is a product of sophisticated teams deftly applying real-time data to their products and processes.
For news, this is obvious. We all get that, especially in a world increasingly obsessed with what’s new, now with a daily opportunity for winners/losers where everyone badly wants to be the former. But real-time use case is across industries — as everyone has events they create, marketing tactics with a temporal nature, and live, active users to their sites and mobile apps - all of which present opportunities for real-time analysis + application. Bottom line is you don’t need to be the WSJ, NYT or a large social media platform to reap the benefits of activating real-time. They’ve honed real-time signals to inform their strategies. You can similarly.
Just a few areas within an organization that benefit from real-time
Before we get into the how, let’s take a look at just a few areas within an organization that benefit from using real-time.
Marketing: Diagnostic/testing of campaigns in live environment, real-time capitalization on trends/events, showcasing success to stakeholders, and personalization/amplification of content.
IT: Diagnostics/troubleshooting of sites, watching real-time in tandem with other metrics such as server load during system testing/optimization, cross-functional monitoring of enterprise team usage of intranet/applications. Analysis of company activity/productivity.
Design/dev: get early, instant data/user behavior on new site or app designs/architecture before rolling out more widely.
Customer service: During a crisis or customer service issue at scale, many monitor use of site content and ensure they have the correct response which is findable and make sure to respond to as many people as possible (or sometimes just the right ones, if resource constrained or the strategy dictates such an approach).
PR/social: Real-time content creation to capitalize on breaking news, events, setting them up for success and being positioned as “go to” voices for reporters actively looking for qualified commentators. You want this to be your company’s subject matter experts/execs as often as possible, and done frequently enough people will start to come to you organically and without even needing to do much at all. Also extremely helpful during crisis situations and approaches here should be in playbooks of every PR team/shop worth their salt.
SREs: Large tech organizations have SREs (site reliability engineers) who actively monitor real-time, especially during anticipated spikes in traffic such as national ad campaigns or new product launches. Coinbase’s site/apps going down during their Super Bowl ad was a huge miss especially since it cost them $7M for the campaign, better marketing/SRE team coordination means preparedness to mitigate outlier events like causing issues before they ever occur (but not without ample x-functional planning - so crucial!).
Their team is already classically known for downtime during high volume days in crypto markets, so this is a bit of an ongoing joke, one in which I’m still vexed by their inability to execute better on.
Seems like there are some (public!) indicators why this might be the case, at least on the marketing side.
Anyway, as you can see, real-time data is one which many areas of an organization may benefit from paying attention to and creating processes for. Above is of course just a shortlist to get you thinking; as always you’ll need to create the appropriate approach for your teams.
Some key basics about real-time for new analysts
Real-time is the ultimate diagnostic tool: Digital marketing is a set of interconnected systems, an engineering problem essentially. There are search engines, web content, social networks, mobile etc, and when you start executing tactics it’s easy to get things wrong, but not always easy to understand what. The ability to see things in real-time can help provide a quick diagnostic view (answers the question: how did the system react to that instantly?). A simple example of this is to make sure campaign tracking is correctly implemented before launching a campaign widely. With real-time reports you can find out in seconds whether you’re getting the data you want in analytics or if there are unforeseen setup/tracking issues.
Real-time provides alerting/intelligence: It can give insight into things that are new or different in addition to taking advantage of one-to-one marketing. The right alerts set up in real-time can give you peace of mind on important events for your business such as server-uptime or crashes. Marketing should care as much as anyone else on the technical side (perhaps more!). You don’t even need to monitor this actively, you can setup rules to send you an immediate push if something isn’t how you expect.
Real-time can let you win social by capitalizing on trending topics: As the web continues to move to focus on streams, understanding how to interpret this data becomes important. Observe, understand, act (i.e., if you observed you were starting to “go hot” on TikTok, Twitter, or Reddit, you could analyze why and react by pouring fuel on the fire by quickly turning on paid tactics or ask influencers to re-share — gaining a multiplier effect. There is a very limited window to accomplish this and agility is key).
Real-time on-display during special events or at offices: Helps add energy/excitement to an office environment and can also act as social proofing to show popularity or marketing success or rally your team. No war room is complete without proper dashboards - and I mean dashboard with actually crucial information to monitor that day/minute/second, not simply pretty graphs.
Additional ways savvy practitioners actively use real-time:
Segment and get more real-time data for specific portions of larger sites.
Double-check new profiles, work faster, and not second guess your work.
Create profiles for specific, short-term promotions that are important to see in real-time.
Create profiles for email or other recurring campaigns you might want to see in real-time.
Test virtual page views/other custom code faster.
Perfect timing on social sharing of content.
Stay on top of the daily trending topics/hashtags/keywords as relevant for your brand/category.
Use real-time data to help inform PR pitches (ability to insert yourself into the news of the day, with a legitimate reason for being).
Put real-time data on-display during in-person events: show things like # of people Tweetings event hashtags per minute to encourage others to join in, publish top Tweets in real-time on screens around the conference, etc. There’s so much you can do here, and executed well it extends the conversation well beyond the event itself and to the broader sector of practitioners who may have wished to attend, couldn’t make it in-person but would still join the conversation from their workstation at home/at the office. Bonus points if you can also live-stream the keynotes/best sessions openly for all. No reason all events can’t be an open experience for a sector as a whole to share together, in real-time, if such wider-exposure is a goal (and it frequently is). Ask Apple or Google why they spend so much time here ensuring that occurs - it’s just so valuable and not much additional investment - reach ~5-6K people in person in 1 building, or also add millions more globally online for not much extra work - up to you. The tech sector comes to a halt during Apple’s Worldwide Developers Conference to watch live and be part of the conversation. It happens with smaller companies all the time. Still don’t get why more brands don’t copy-paste the playbook.
Just a few key metrics to look at in real-time (and questions they answer):
Visitors on site right now (new, returning — get a quick glimpse of who’s on your site right now).
Conversions in real-time (are visitors acting on an immediate offer?).
Top Referral sources (what sites are sending my traffic right now?).
Top Active Pages (page/number of active visitors per page — which pages are hottest?).
Top Social Traffic (source/number of active visitors — is social working for me?).
Top Keywords (keyword/active visitors — how are my organic and paid search campaigns working?).
Top Locations (country/active visitors — where are my visitors coming from right now?).
Any outliers in unexpected traffic to certain parts of your site on a given day to dig into/understand what’s going on.
Social/outpost data on sites and communities you contribute to that may not give full analytics, but still provide useful information such as trending topics.
Now that we have a better understand of some areas of real-time pro marketers and analysts think about, let’s go through additional high value ideas and examples to get you thinking even more.
Ways to use data to improve your marketing in real-time, while you sleep
Putting on your analyst hat, conducting analysis of your site or app, and fleshing out recommendations is just the beginning of making data-driven decisions, especially with real-time. Once you’re in a groove with optimizations, the next step is to put data to work for you, to improve the customer experience and increase conversions while you sleep. Especially since it’s not humanly possible for your team to sit at screens 24/7. Always work hard to free them up for the important work.
Let’s go through three areas you should think about programmatically using data to improve your marketing in real-time along with some additional tangible examples. When I say programmatically, in plain English, all that means is letting machines do the work and put your data to work for you, productively.
Use real-time analytics to dynamically improve your site content, and conversions
Most popular brands always have visitors on their site. While there, these visitors are viewing content, sharing things they find interesting with their social networks or via “dark social” such as texting/Discord, making purchases and more. But are you using these actions to improve the experience of other users?
Looking at this data in real-time and taking action on it is not always feasible. But, using a Real-Time API to dynamically improve your site content is. Using an API from your analytics provider, you can make queries about your real-time data and use that information to create a more compelling experience.
One of the most basic but practical use cases is to manage the content on your webpage. For example, you can query the API for the top visited URLs to construct a “top trending content widget” and display the number of active readers. This can help create a greater sense of urgency to view and share ‘hot’ content (as I mentioned above media companies do this all the time, and achieve great results - there’s a reason this exists on nearly every media company site/app).
Additionally, this metric can be shown on different conversion pages of a website to impart a sense of urgency and demonstrate demand for a given product. Twiddy, a family-owned vacation rental company has previously tested this. Larger brands always have a call to action or incentive, and using the Real Time API allowed Twiddy, a smaller player, to have the capabilities to compete. They were able to capitalize on traffic to the website, since showing how many other folks are looking at a similar search shows the demand stream. This creates a sense of urgency, similar to an online auction. Not only did their revenue increase 18.6 percent, but the average order value increased 11.9 percent and the conversion rate increased 7.9 percent - in the slowest booking period of the season and also during the slowest demand period of the year.
See the Twiddy case study for the full story and the screenshot below for an example of how this looks visually on their search results page to show a virtual representation of real user demand to persuade users to book that rental now, before it’s too late:
Use your own site analytics to power remarketing in real-time
Remarketing is a key tool to increase missed conversions and reduces impact of shopping cart abandonment.
Ever visited a website intending to make a purchase but got side-tracked halfway through with another task? Of course you have. It happens to nearly everyone. The good news for marketers is you can ensure these users get a friendly, non-invasive follow-up to help remind them to purchase (convert). Imagine you are browsing 1800flowers.com with internet of purchasing your mother something special on Mother’s Day, get sidetracked with work and forget. Remarketing is a tool that can help a brand manager automatically save the day and ensure mom gets her flowers. This is a true story, as it happened to me during a previous Mother’s day - I even randomly met their digital team during a conference a few years ago and got to tell them in person how much I admired their work, always fun.
Remarketing with Analytics simply and efficiently lets you follow up with people who have already visited your website, and deliver ad content specifically targeted to the interests they expressed during those previous visits using signals you define. If you’re new to remarketing, my former team at Google produced an introductory Webinar to help you get started. There’s also a simple and actionable “how to” guide from my friends at world-renowned digital analytics consultancy Cardinal Path to help out.
Run tests that automatically predict/pick winners
Your landing page is performing well, but you can’t help but think it could do better. You could replace it with a new design, but how do you know if it’s truly a better experience? Worse, what if it performs poorly? You’ll lose revenue and conversions while you’re busy running the numbers. There’s a better way: use a system to test your content using real, live traffic as efficiently as possible, find the winner and allocate 100% of visitors there ASAP.
If you’re a Google Analytics user, Google Optimize (see example above) allows you to compare how different pages perform using a random sample set, define what percentage of visitors are included in an experiment, test different objectives and get automated updates on how experiments are doing. And the best part about the system is it works to get a statistically relevant sample set in many cases much faster than you think to minimize the number of users who see a poorer performing page.
Regardless of which tools you use, the point is that you can leave the hard work to software to decide how to smartly funnel users to specific tests in real-time to get to winning results fastest, and free up your time for more strategic and creative work that only humans can do. Some people have started to call this “growth hacking” (I despise this term) or use other buzzwords to describe the process of automating parts of your marketing with data, though it’s not new and actually doesn’t require as much technical knowledge as you may think. Buzzwords and fancy jargon aside, this type of optimization tactic as part of your marketing roadmap the future of smart marketing.
Quick wrap
The web (and of course mobile) is getting faster, and not just the speed of the pages, but also the speed of change. Before, it was fine to build a website or mobile app and modify it only when new products were launched. But all of us analytics nerds know that’s just not enough. We need to be constantly on the lookout for problems and opportunities as they happen, and where appropriate use data to solve them. This is something real-time can efficiently enable, used wisely. I hope today’s post will encourage you to think about best using real-time as part of your team’s marketing and analytics processes to solve problems unique to your company.