E-mail is now a fundamental driver of many new business initiatives. But audience data, including e-mail addresses, is typically stored in a chaotic array of separate systems and platforms.
To make sense of - and leverage - this data, a best practice is to combine e-mail data into one big database, which is a complex, time-consuming project. But since this data is accumulative, an early start provides a significant competitive advantage. Ultimately the "e-mail database" is really much more; it is a highly detailed database of customers. Here are nine planning steps that will set your company on the right course.1. Identify objectives
What are the anticipated objectives for the database? Clear and concrete objectives drive execution and make it simpler. Data accumulation takes time but having clarity on what products on the road map will be driven by e-mail is a critical first step.
Some companies have limited, immediate objectives, such as capturing every e-mail possible to drive group deals.
However most local media should have additional long term objectives, such as building profiles on these 'customers' so that the data can increasingly be segmented by age, gender, zip code and area of interest. Wiill your company want to sell targeted e-mail blasts? Create newsletters that drive other initiatives, such as SMS sign-ups or coupons?
What kinds of promotions are advertisers asking for?
What other projects are on the road map that will require promotions?
What about marketplaces that will be driven by targeted e-mails?
How valuable is it for advertisers to be able to include an e-blast to the correct demographic in the correct zip code, as part of their promotional campaign?
Owning the best consumer data in your local market requires a database that accumulates knowledge on consumers such as zip code, age and gender, or take actions such purchasing a ticket to an event.
(One concept Groupon brought to the table is that local media visitors are not just an audience, they are customers and will respond directly from local media and partnering merchants.)
Both the head of interactive and the head of digital revenues should participate in meetings about key objectives and what data will be accumulated, since they are ooking at product development on one hand, and what advertisers want on the other hand. Consider having data managers report to the executive responsible for the resulting revenue streams; the key objectives for data acquisition and management will become immediately clear.
2. Inventory All Lists
Make a list of all the existing e-mail or other audience lists in each market. Then for each list, list the type of information captured, what systems the information is housed in, the number of records, if it can be exported, how many segments there are and so on. Looking at these lists determines the critical question: How big of a project is this?
Sources are important to creating segments so always preserve the source. People who signed up for a free pizza from an advertiser's mobile offer, for example, can be combined with those who checked a "please send me dining offers" box in a field tagged "dining." So preserve the source on a go-forward basis.
Identify categories of data, such as an interest in dining, that are already indicated by the data, even if limited, and give it these separate fields. In addition to age, gender and so on, interests will be valuable to the profile.
This is also a good time to categorize levels of relationship. Basic kinds of relationships require different definitions. You can start with a few simple ones:
a. Cold - The list was purchased or traded and has no relationship. These lists are used to create customers (as defined below) and to infill marketing for large advertisers who want more coverage than your list can provide, say, in specific zip codes;
b. Customer - The user has converted to something, whether purchasing a subscription, signing up for an advertiser's promotion or registering to receive site news and
c. Opt-in - The user has specifically opted in for delivery of news, offers, etc. and therefore will sent only offers that apply to its category or tag. This is important to keep track of, since they are the most sensitive to spam, which is anything they did not specifically sign up for. Does this cover it, or are their more to include?
Databases in an of themselves are usually simple constructions, but combining usually raises a number of thorny issues.
For example, the circulation database may not have email or cell phone numbers. So just to get up to speed may require filling in this important piece of data before the asset can be fully leveraged. We've also seen circulation data that doesn't export. Yikes....Good to know. Or data that has been preserved in e-mail management systems but which no longer have anything like a source and is largely undifferentiated except for open and conversion rates. Or, conversely, a list may have a variety of information that is not relevant today, given that none of the other data will have these fields. Should you keep it?
After recording the inventory, you should be able to create some project phases: ie what data and data fields the company will be able consolidate now, and what actions are need to fill in data that is too incomplete or otherwise unaccessible.
Note: Don't forget to include the database of advertisers - amazingly, most local media companies do not do have complete e-mails for advertisers. Advertisers are people too, can be included as local consumers, and are often impressed by seeing your marketing at work.
3. Architect the database
This requires looking at the list above, and deciding what fields will go in the new central database, what kinds of sorts and filters will be needed.
So from the inventory, create a new list of fields, sources, categories and relationships, as well as zip code, age, gender.
The number of fields and filters that will be needed now and in the future will help determine the choice of platform.
4. Choose the platform
Once you have the requirements, it's time to look at database platforms.
Reversing the order of this process traps the data into a system that may be insufficient. So, when looking at both email and database providers - the two are related since they will need to update eachother - make sure that the fundamental data can be stored, and filtered, as well as being able to create unique audiences and capture conversions.
One sina qua non criteria is that there must be a way to filter out opt-in data so that this audience will NOT be sent additional promotions outside their specified interest. That is typically difficult without some kind of way to sort via multiple variables, ie a promotion for a health club that goes to all women registered on the site EXCEPT those who opted-in to receive only dining news.
You can see how the capabilities of both excel and less flexible e-mail systems are quickly exceeded.
The biggest decision will be whether to use - or change in order to use - an email provider to host the data, or convert to a more flexible database such as SQL or a proprietary system. Does the database provider have enough fields and filtering options? Will email conversions be incorporated back into the data and if so how?
Data-mining email conversions is also a source of data which is usually secondary to demographics, but more pure in terms of known responses to specific promotions, so accumulating all of it in a central database is incredibly valuable.
Sometimes the value of segmenting is being able to filter out segments of audience who are not closely related, not because they didn't opt-in, but because they can then be sent other more relevant promotions. In other words, if you have eight promotions going on, you may not want to send all eight to the whole database and burn them out. You want to have, ideally, eight segmented audiences. So segment your audiences. You can't pull the rabbit out of the hat unless you put the rabbit in it.
So what platform to use? Excel cannot easily filter more than one variable and becomes unwieldly with too much data. It is the best platform for exporting and cleaning data (ie renaming and reorganizing fields).
Likewise, most e-mail systems that may excellent at reporting conversions are not designed to filter a deep level of other variables. MailChimp is a recommended solution for small companies, since it is highly flexible and can support a number of data fields, such as age, income, zip code, and so on, to create custom audiences. But larger companies will outgrow this system.
Some argue that the data should be housedi n a CRM system, such as SalesForce. Considered overkill for small ad sales departments, this platform does, however, have both a CRM system and e-mail function. The functionality of the e-mail system is a core issue in taking this route.
Very large companies may want to build an SQL system that is open source and scales, but will have to work out the interface with email.
The choice of database platforms depends upon both the size of the company and its objectives. For under a 100,000 data-points, and without full profiles, a robust e-mail delivery system like MailChimp may work. For over 100,000, a custom solution, a more robust database will be neccesary.
5. Start with excel
No matter what database platform is used, the starting point will be exporting to a simple CSV excel file, since most data will be imported from these files anyway. Many systems will continue to use an Excel database to clean and add data and as a back-up.
This data can then be imported into a universal database, or, for smaller companies, the e-mail or a CRM database.
6. Map existing data fields
As the company starts exporting data from multiple sources into Excel, there are bound to be different labels for different fields. Make sure someone looks at each database that will be migrating into the big database, and renames it so that each field of data ties to a field in the new database. This chart (ie any renamed field that is not immediately intuitive) should be noted some place on the spread sheet document, so that it is available to anyone else who may work on the database in the future.
In large companies, a list of the nomenclature and definitions should be routed to data managers throughout the company.
7. Import and de-dupe the new database
Any mature media company will have duplicate data (for example advertisers who also subscribe to the publication, or subscribers also receive an email newsletter). How will these combine to eliminate duplications, especially if there are different spellings of names?
Since you don't want the best data to be eliminated in the duplication check, develop a hierarchy of lists – newest to oldest, email list vs circulation - with the goal of having the cleanest, most recent data at the top and merging older data into it, so that the best data is kept when duplications are eliminated.
8. Create a job position to manage the data going forward
Depending on the goals established at the beginning of this project, someone in the company needs to be responsible for data maintenance and integrity. Data gets old fast. Make sure to put in place a plan to refresh the data on a regular basis. Peter Konti, of Kelsey Group, recommends a 20 hour a week position for a mid-sized newspaper, so labor resources can be roughly extrapolated from that figure. Our recommendation is that either the key revenue officer on the digital side, or the head of interactive be at the top of the reporting chain for database management.
9. Keep it Simple
If combining “all” the data into one database is not immediately feasible because of size and complexity, break the project in into smaller chunks. The first breakdown, if there are multiple markets, should be by market, keeping a consistent nomenclature and organization of databases between markets. Within individual markets, look at simple solutions first, such as combining a current circulation list with registered users on the website. Print companies have the advantage of large circulation lists to work from and a real shot at owning the most complete consumer data in their markets. Starting to accumulate data now will create powerhouses five years down the line.
In making these decisions, refer back to the goals, and what data will ROI best and first will usually be the top priority.
10. Share information
Established media are woefully behind at organizing all the data they have and utilizing ways to manage customer relationships electronically, including advertisers, much less consumers. But the sooner you start, the quicker this skill set will become ingrained in the organization. So for small media companies, create a small group of like-minded executives in non-competing market who are committed to embarking on ownership of consumer data in their market, and share your RFP's for data management platforms, nomenclature of fields and so on. If you are interested in creating a group we can help. Please contact us at firstname.lastname@example.org or email@example.com. For a list of eight ways to build e-mail lists click here.
Jane Bogue, who collaborated on this article, is an independent digital media consultant out of Wells ME. Her company, Wilbur’s Point Media, develops digital sales training and revenue strategies for traditional and digital publishers. She can be reached at firstname.lastname@example.org or 207-337-4313.
The author, Alisa Cromer is publisher of a variety of online media, including LocalMediaInsider and MediaExecsTech, developed while on a fellowship with the Reynolds Journalism Institute and which has evolved into a leading marketing company for media technology start-ups. In 2017 she founded Worldstir.com, an online magazine, to showcases perspectives from around the world on new topic each month, translated from and to the top five languages in the world.
No comments on this item Please log in to comment by clicking here