Too Much Data, Too Little Insight
ADOTAS – Digital marketers are inundated with data, yet struggle to quantify program performance. This seeming contradiction — tons of data but only a rudimentary understanding of effectiveness — is prevalent across the digital landscape. However, there are solutions to help rise from the tsunami of data with insightful knowledge in-hand, and seize business opportunities previously not visible.
“Figures often beguile me,” Mark Twain wrote over 100 years ago, but it expresses the frustration and analysis paralysis of many digital marketers today. Why is measurement so difficult when we employ armies of brilliant analytics people using extremely sophisticated tools? There are three primary obstacles:
1. Too Much Information (TMI). The web is so data-heavy, it’s easy to miss what’s really happening. Numbers may not lie, but they can mislead if you aren’t looking at the right ones. Every day, digital marketers wade through immense volumes of marketing program data, scouring reports of clickthroughs, open rates, costs per action, or new engagement statistics, most of which don’t answer the core question about what a marketing initiative is contributing to your business.
2. Unmeasured objectives. Too often marketers haven’t mapped out specific digital metrics that align with their business objectives. When you can’t connect the dots of marketing activity to goals, you don’t know what to look for, so you look at everything, crunching excess data and reporting too many metrics. Measurement should indicate progress toward business goals, and point out actions to help achieve them.
3. ROI overemphasis. ROI is revered as measurement’s Holy Grail because it draws a direct line from investment to impact. In practice, calculating ROI is sometimes impossible – information is unavailable, consumer behavior is too complex, multiple influences can’t be separated, or costs aren’t properly allocated. Focusing solely on ROI drives a cycle of unrealistic measurement expectations and disappointment. Tactical performance metrics are also needed to guide improvement, with or without ROI.
“Not everything that can be counted counts, and not everything that counts can be counted.” — Albert Einstein
How do you transform raw data into knowledge and insights that identify business opportunities? As digital measurement evolves, here are some immediate steps to help:
Focus on the critical few, not the insignificant many. Focus, focus, focus on the few metrics that matter most. Just because you can, doesn’t mean you should, measure and report everything. Take your top objectives and assign 1-2 priority metrics against each. Take a reductionist approach to marketing analytics, prioritizing valuable information, and separating measurement wheat from the chaff.
Make current metrics more precise. Some metrics are simply inaccurate. Attributing actions to the last click over-credits Paid Search, and ignores display’s impact (22% lift on search conversions — Atlas), so ascribe some assist to all marketing exposures leading to a final action. Engagement must be specifically defined for your business, because if your site goal is easy registration or checkout, Time Spent on Site is the wrong engagement definition, although it might be perfect for a brand site.
Avoid direct response only. Action metrics speak loudest, so 30% of total direct response dollars are spent online, but over 90% of branding spend is offline. Digital rushed to deliver direct response and sales-attributed metrics, but largely ignored impact on awareness, attitudes, and branding. It’s time to avoid solely focusing on actions like the click, and consistently incorporate brand impact, qualitative site feedback, and content engagement metrics.
Identify proxies. You often can’t track ROI or the direct path to a sale, but can measure behavior that’s strongly connected. One marketer conducted site visitor research and discovered that a specific download was highly correlated with a sale. They then reoriented their marketing to drive that download and measure success against this new-found proxy.
Continuously learn and improve. Structure your programs to monitor success constantly — from launch, through implementation and afterward. Set aside 10% of your budget to test potential enhancements, so you can quickly deploy all effort only to initiatives that work best. Some call this “failing immediately” because you almost instantly know what isn’t performing and eliminate it.
Assess all key initiatives. Take a comprehensive measurement view, establishing at least one key metric for each area of focus. Define success metrics against each business objective, marketing tactic, and key target or constituency. Spell out top metrics for each stage of building a customer relationship — Awareness, Engagement, Action and Retention. Incorporate all metrics categories — qualitative, quantitative, financial and customer profiles.
The pressure to measure and the volume of data sometimes seem overwhelming. Do your best to calculate ROI, but don’t ignore other key metrics that indicate whether your marketing programs are performing successfully against your business objectives.
This is a great article for the simple fact that a shift how we think about online advertising is encouraged.
I believe that” Action metrics speak loudest, so 30% of total direct response dollars are spent online, but over 90% of branding spend is offline, ” is fundamentally true, and this quest for the dangling carrot of definitive ROI has increased the amount of data collected on campaigns to the point of white noise. But why did this happen online in the first place?
The foundation of ad commerce has always been viewership. A publisher sells his best estimates of how many times an ad will be exposed to a viewer for any given ad buy. Of course other data related to ROI is important but it all stems from this core. Once you know your viewership, you can extrapolate ROI from there.
CPM Impressions, as counted by server log file based ad servers today , offer CPM counts that are so inaccurate (up 95% never appear) , online advertising has been forced to the “kids table” of advertising mediums and a mess has been building on that table for years as new data is layered on. Branding clients have long been aware of this issue. This I believe is why 90% branding spends are offline and so much ancillary data is collected; to offset the inaccuracies of the server log file CPM metric.
There is a new “Viewable Impression” metric (Called RealVu) now available that is Media Rating Council accredited and IAB compliant. This metric counts and impression only when an ad is actually visible to a viewer.
I believe that until the new “Viewable Impression” metric becomes used ubiquitous, branding dollars will remain “off-line” and “too much data will too little insight” will prevail.
Thanks for getting out there and calling attention to a shift that needs to take place for online advertising.