Thoughts on creating a coherent data strategy
Websites, whether they’re large premium publishers or small, niche players, will want to think about the difference between defense and offense when charting out a data strategy. Defense is about protecting data generated by consumer web interactions. Offense is about harnessing that value to drive new value from content, commerce, collaboration, or advertising.
Defensive data strategies center on
- measuring the volume of data collection by external entities by site, category, and time and its impact on web page performance
- understanding how third-parties chain themselves together to mine data from a website
- focusing organizational attention on the potential opportunity loss that results from uncapped data leakage
- reconfiguring T’s and C’s with agencies and advertisers to establish new norms and thresholds for acceptable data collection
- applying new methods to block unwanted data collection by incorrigible offenders
Data offense, meanwhile, is about claiming value from data. Offensive data strategies center on
- enriching the web operator’s understanding of audience via detailed, flexible analysis,
- creating and managing audience segments dynamically, drawing on both first-party data and third-party sources, and
- improving revenue performance by connecting those segments to advertising, content, or commerce opportunities.
As web publishers establish a business plan for data, they will find it productive to organize their efforts around Sources and Uses of data.
Data sources could include:
- First-party web operations, including information inferred or observed during users’ web visits (e.g., attributes inferred from IP address, content consumption patterns, referral keywords, site searches, etc.)
- First-party performance tracking, drawn from ad serving or similar systems, recording user actions tied to content, commerce, or advertising (e.g., clicking on a call to action, or converting through some sign-up, download, or purchase). These could be specific, such as a user’s response to a particular cost-per-click advertising campaign, or general, such as a broad categorization of ‘users that click.’
- First-party registration and subscription data, volunteered by the user as a part of their web experience (e.g., pay wall content subscriptions, newsletter sign-ups, app downloads, purchases, etc.)
- Third-party datacollected from online and offline sources and aggregated for resale by pure-play data providers (e.g., basic demographic data as might be available from TARGUSinfo, credit ratings from Experian, or purchase intent data from Exelate)
Web publishers must evaluate all of the potential and likely uses of audience data in support of their content, commerce, and advertising operations.
Data uses could include:
- Enhance value/increase sales of existing products– use audience data to capture more value from existing products monetized at by higher prices, and/or to drive higher sell-through and share-of-spend by making existing products more relevant and compelling to existing customers
- Develop products to capture new demand– use audience targeting to characterize users and advertising to tap into new sources of demand and incremental revenue
- Increase salable inventory by finding audiences outside first-party web properties– use audience data to increase reach and frequency of audience inventory at secondary premium prices, finding certain segments of users elsewhere by targeting those users via peer-to-peer cooperatives or by bidding on user-targeted inventory from real-time media networks, resellers, or exchanges
- Develop data only revenue streams – for torso and tail publishers especially, identify salable audience assets and take those assets to market via private data storefronts, data marketplaces and exchanges, or private opt-in media/data cooperatives with trusted buy- or sell-side partners. Such standalone data assets could be unique and high value (e.g., purchase intent data from an auto content site) or more commoditized audience characteristics (e.g., basic non-PII demographic facts collected during a site registration process).
- Increase user engagement through targeted content– analyze user behaviors, preferences, and intent and using that learning to customize content offerings, leading to stickier audience relationships, more efficient content discovery, richer user profiles, and more salable inventory
- Increase transaction revenue through targeted commerce activity– analyze user behaviors, preferences, and intent, and use that learning to customize commerce offerings and improve conversion funnel and sales of goods and services
These distinctions suggest the following decision inventory to help the data strategist complete her data strategy:
The preceding is by no means comprehensive, but we hope it is helpful by way of catalyzing productive dialogue as web publishers work to flesh out their data strategies. If you have questions or ideas, I invite you to weigh in.