Adventures in Mobile Marketing

Geographic email programs are very, very successful. They are difficult to get, as email marketers, since we can’t determine locale from email address. I’ve had discussions on discussions about matching an email list with ISP and there fore region, but the false-positives outweight the real positives. The old saw is, the massive amount of Vienna, Virginia folks who are AOL subscribers (and do not reside in Vienna). First, though, why is it worth the trouble, and lastly, how to determine geographical location.

What rocks about regional campaigns

As usual, don’t believe me, do your own test. Take a randomly selected control group, and then a targeted local group, and serve local content, and serve diffuse general content. You will see the response rate- opens, clicks, purchases- skyrocket in regional campaigns. Lately the Obama campaign has been regionally SMS’ing and emailing, anything from “call undecided states,” to “meet-up at Temple bar for the debates.” Past clients have further created regional relationships with other businesses to bring potential customers in the door with discounts, or invitationals.

I think regional works so much better than non-regional campaigns because, as consumers, we already have a host of relationships, impressions, and brand recognition with those regional businesses and locations. So when a new brand makes an association they can leverage off of all of those positive feelings, and one major feeling is trust. As all CRM customers know, if you can bring more trust into the relationship you are farther towards a successful one.

How to get Region?
This is tough data to get for an email marketer. Various methods of receiving region:
- Phone number is regionally located (see Obama SMS campaign for getting mine)
- Positive activity on regionally focused emails – you can provide an email campaign with a list of regions, and whoever clicks on X region can generally be considered interested in that region. Note, your email system has to be sophisticated enough to segment on clicked links.
- Asking for zip code (generally considered by consumers as less identifiable and easier to get) in a contest or just outright, for regional email messages
- Changing the online sign-up or registration information to get this initially

Last tip: Please put the region in the subject line. Show off that you know it, and that this email is tailored to the recipient!

Examples:
Williams-Sonoma. They have targeted me as San Francisco Bay Area, but they neglected to put it in the subject line, which have really pushed me to open it.

Uploaded with plasq‘s Skitch!

Of course regional campaigns are a travel company’s baileywick, but even here they do a kind of silly goof:

Uploaded with plasq‘s Skitch!

So the general idea is that you have a customer database, and you feel it’s inadequate so you go to one of these companies and send them info, and they match it and send it back to you. Sometimes you get these cute scores and nicknames based on zip code, like mine, “Bohemian Mix.” I’m supposed to have an Audi, read the New Yorker, etc. None of that is true.

Managers who decide to buy data appends are crying out in desperation for insights into certain hidden segments of their database. What they really need is an analytical tool to help them dive into their own data. Acquired data is not going to tell them anything useful. Instead, they have a lot of information, they just don’t have the tools to get to it.

Three main reasons why I avoid them at all costs- is that they require you to submit data which they then add to their databases and sell the findings to other customers (anonymous-ly, but still). More importantly, there is no visibility into the logic that they use to determine models and segmentation, and neither are the sources. Also, it’s largely without consumer permission.

But companies have permission from their customers- from interactions. If the companies just take the time and effort to dive into their own data, they can find out trends and behaviors far more significant and relevant than those provided by these vendors.

Try it at home- as I’ve done on numerous occasions to prove the uselessness of these farms- randomly split a list into a control, and two test groups. In group A, further segment using your own behavioral thresholds, in group B, apply the relevant data farm’d segments (i.e. “Bohemian Mix,” “Digerati,” etc.). Give the control group a randomly pruned list of the same size. My experience, doing this on various campaigns over a year, was that data farm segments provided less or no lift than internally derived findings on purchase history, interactions, and past campaign behavior logic, to name a few.

I believe the savvy email marketer has to recognize (and avoid) their own bias, rely on previous findings from tests and analysis, and continue to test and model to create an ongoing, maturing view of the customer. Quick fixes like data appends simply muddy the waters, and worse, substantiate assumptions about the customers without data, testing, and findings behind it. Their sources are too diffuse, and the wins aren’t big enough.

Another issue for me is that to use the appends you must provide them with data- this in turn creates the very product they are selling back to you. Repackaging data and forwarding it on is unethical in my view, a disservice to the consumer, and costly.

A former colleague chatted me yesterday with this perplexing “deal”:

SalesForce is offering non-profits free licenses to their software. But we keep on running into issues with their “governors.”

That’s really great of SF to do- and explains an uptick in “do you do SalesForce?” emails that I’ve been getting. The downside is a serious throttling to list size. SalesForce “governors” control the amount of items in targeted lists, so you can only include very small mailings. Seemed like all of the downside of having a hosted solution with none of the upside. If your experience is different, please post here, my friend’s eager to have an alternative, or fix, to the solution.

The question came up of complex segmentation for small lists. What ESPs can handle it? We determined that manually building a statistical model and scoring it then uploading to ESPs and using the “score” to simple filter the lists, is probably the only real inexpensive solution for non-profits. Note: this is not for transactional, or triggered, messages, just promotional and newsletter-oriented emails.

The alternates I suggested to him: SugarCRM, StrongMail, and a few consultancies that could help him setup the database. He and I had worked on an Epiphany installation and the license for that was too expensive for his clients, so that was out. He wasn’t interested in a per-email pricing either.

I stayed at a Marriott recently, and it took me half of a morning to find my Rewards number, so they could assign my points. Sure, it’s my fault I didn’t carry around my loyalty number around with me- but we all know there are so many programs it’s near impossible to do that. It’s almost expected now that nobody will have their member card with them. Barnes & Noble looks it up by last name and phone number. My video store, the same.

I stop by the front desk on the way to breakfast. I don’t have my number, so they lookup my Marriott rewards account by my last name and the billing zip code from the first time I had registered with them. I’ve only had two spanning the last 10 years, but no, neither work.

I remembered I had an email from them on my iPhone. I scrolled to it in front of the registration person and no, nowhere on it was my number. My points, my first name, but not my member number. (see below)

I tell the registration person not to worry and later that day when I remembered, I logged into the site. Still not so easy- you have to click around to your account, profile details (it’s not on there), and then finally the account profile (what’s the difference?) and there is the number, hidden amidst other details. The reason my zip code didn’t work is that it requires the four digit extension. I wonder if they require the dash?

Later that night return to the hotel and provide the number, and then they assigned points. I doubt my parents, or colleagues, would be as determined as me to get these free points. So the kind of rewards member they’re targeting is the cheap, coupon-clipping, business traveler. I doubt that’s their desirable target demographic.

On the site, if you forget your password- and if you forgot your number, you probably forgot your password, is accessible by filling out this strangely lengthy form:

I finally found it because I punched through my usual passwords and struck gold. The lessons I take from this: if someone is already a guest, and they want to assign points to their loyalty program, they should be searchable some of the common criteria .
- email address and/or
- first and last name
Also, the email should either have the account number in the email, or at least, one click away. Currently it’s 3 clicks, login, account page, then profile.

I wonder if they are trying to get regional data from their customers, because hospitality really leans on regional information. The problem is that they are creating this barrier to participation in the program. Perhaps that’s intentional too. If so, this is one of the most intricate, and relatively useless, process in CRM. We build this system, but don’t really want you to participate. We make it hard, so you won’t really take advantage of our offers.

I know it should be different online where there are more fraudulent schemes, but doubting or questioning someone who is already paying X dollars a night in the hotel seems contrary to the hospitality industry’s general attitude towards the customer.

I wonder at these loyalty programs- you get a great segment of your customer database, but is it worth setting up an entirely different experience, and managing the data, which can go awry like in the example above? WellsFargo has a very easy Rewards system that happens very passively from the consumer point of view, no unique number, just a link you click on to agree to the terms, and then you start accruing points.

Just when I thought this conference (MarketingSherpa Email Summit) was too lightweight, I stepped into a great talk- a little misnamed, but regardless- very interesting. Matt Griffin, director, sales & marketing of the Boston Celtics, talked about targeted messages and a series, multi-channel campaign that really has no need for improvement (despite his own admission, see end of post).

Roughly, they have 3500 unsold tickets and have 30 days until the game commenced. They designed a series of emails to sell through in aggressive campaign.

Some high level marketing aspects of the Celtics: they have a regional marketing focus– around New England, 75 mile radius around stadium. Their profit, as tenants, is mainly around perishable, event-based items (tickets) and sponsorships- nothing else, no parking or concessions. This is a major reason why pushing to sell out games is so important. 80% of the tickets are sold online, which is quite a flip from 10 yrs ago, when it was 20%. They focus on the impulse buy.

Some more breakdowns, just in a bullet form, for their marketing, and knowledge of their customers:

• Very emotional customer target/segment
• Very broad demographic appeal
• Varied reasons customers come to the game
• Pre-defined customer clusters, clickthrough behavior is segmented out, pre-defined
• 800,000 price points – every seat has a different value that fluctuates day to day, email marketing tests and probes price tests
• Reacting quickly (from the click traffic) helps with the quickly shifting win/lose season with the team

I asked him about static segmentation versus dynamic. I wouldn’t think pre-defined was something to brag about. I pride one of my clients on their ability to shift around their segmentation as much as they want. He said that they shift it around every year based on statistical reports and that is usually enough, he’s found.

The case study he talked about was the 12/15/07 game against Denver Nuggets, holiday season meant not a full house, so they identified the open seats, which were in the upper balcony/loge, and then designed a 2-phase multichannel email campaign.

First phase was a targeted email to pre-determined clusters, to sell the large blocks & balcony, which are great for a group sale. Various sampling of targets were:

• Every customer who had bought on same night as this game, Friday, which is basing it on purchase time and date
• College domains
• Special family offer to customers with children- Offer was hot dog + soda, drop price, create a package,
• Anyone in past who had bought,
• Target against key players’ alma maters such as Syracuse, alma mater of Carmelo, U-Mass, which is the alma mater of Marcus Canby, etc.
• Survey takers got an offer, contact-us offer ticket;
• And almost all service outbound tickets had specific Denver offers.

All in all, there were 24-36 distinct offers created for filling the stadium for this Denver game, most based on the segment and data points for each demographic.

Phase 2: Clicker conversions, multichannel. They generate a a “click list”- all of the customers that had clicked on links regarding the game, but had not purchased, and issue that list to telemarketing group. These are highly effective list calls, the sales people really like it, salesperson has the entire list available to them, on the phone, they can re-offer better offers or seats, or determine reason and follow-up or store the reason for refusal. ROI is better than email on follow-up phone calls. I asked him whether they store those responses for analysis, he answered they do, and I think that is a real key to evaluating offers and campaigns. Very impressive data capture.

The result of this effort? It was a sell-out game, Pierce told the marketing group that “the crowd really carried us,” which is something “us marketing folks feel good abou.” Score: 119-114, and it was the best game of year for this team.

Phase 3, was to follow up on the attendees, selling 6-game package, season passes and merchandise.

These series generated over 2x over standard email campaigns. Because this kind of sales drive was a near sellout, the initiative has become standard with for games with same ticket buying standards. In the league they are the best practice at doing this, Cllick Tactics is the partner.

Questions from the audience…

Q: How do you manage contact frequency, is hitting them every game? Also merchandising, surprised you weren’t doing that earlier.

A: how do we avoid email fatigue? only send out 1/4 the database, and rotate, so not more htan 2 a month from us (wow, very infrequent) learning from 2 years ago, when we “carpet-bombed”, and our consumers were getting to coupon-hunters, offer-hunting, so not conditioned to keep waiting for them

Q:Why are you not focusing on merchandise, thought that would have been done already.
A: We’re trying to get premium packages and merchandise, jersey, playoff-strip, trying to get more than face-value.

Q: Clicker-non-converters, how do you follow up with them, if you do?
A: People who click and don’t purchase, less effective leads than non-click/non-purchase leads, strangely enough. I’ve tested, and that’s how it turned out.

Q: (my questions!) Do you store the refusals from telemarketing? What CRM package do you use? And, why not do dynamic segmentation?
a: Yes, we store the refusal in database. We found that we don’t have to analyze that frequently, once a year is enough. And Archtics (from TicketMaster) is the CRM package (from the ticketing sales system).

Q: I’m a season holder, and very happy with the programs.
A: You shouldn’t have received any emails last year, we suppressed you!

In application and system architecture there’s this concept of generations. Is your CRM system a 2nd, third or fourth generation? Generations generally refer to the sophistication of tools, networked systems, and perhaps size. If a CRM group regularly segments its lists, analyzes those segments, and leverages segmentation in each email it launches (transactional as well as promotional), to me, that is a sophisticated, advanced generationally, CRM system.

Direct mail & catalog marketers have been doing this for a long time- even including more sophisticated scoring and modeling abilities in their lists. One reason I can think of- why direct mail techniques have not been applied full sail to email marketing- is that these are intensive efforts on rapidly changing data, that iterates sometimes twice in one day. Direct mail doesn’t work on that tight a time frame, so they can usually leisurely analyze and segment. Yet there are tools that segment email lists, and the sooner email marketers leverage these tools, the more they will see response rates go up, with smaller email lists that are far more targeted to interested audiences. First, let’s outline three areas that are often confused (I’ve found), and they also mark the technological progression from simple filtering to complex segmentation.

Dynamic content: the template swaps content based on either simple stated logic or static values coming from the email list: for example, product name, which swaps in the hero product photo for each email going out of a specific campaign. The value “pants,” when associated with an individual email record, will tell the template to display a photo of pants. If there is no value, perhaps a default promo product displays, or if “shirt” is in the record, the template knows to display a shirt. Sometimes they can handle a simple is/is not logic too such as a filter.

Filters: Inclusive or exclusive, criteria which determines the final targeted list. You can make these statements complex, “Has not bought, last twelve months, and never bought pants, and has not been cross-sold,” by including all of the logic in one long statement. Each bucket is trimmed by the filter above it. This kind of filtering is limiting, because for every permutation of each criteria, you would have to create an entirely new list. For 3 statements above, that means 9 filter statements need to be written, tested, and output. As you can see it can get complicated, and to recreate this list for each campaign, perhaps they are triggered, transactional emails, then it becomes very burdensome. So we introduce:

Segmentation: Ability to handle multiple logical statements, with a “branching” quality, creates a targeted list with n-dimensional complexity. For example, take the logic above, 3 statements, 3 states of customers. If you imagine it as a matrix, each combination of each state results in a “cell” of the customer base.

The main benefit comes from the ease with which the marketer creates it. Going in and editing the filters, re-using the segmentation in another campaign, and using more complex (than binary) logic, reveals the deep ability to find hidden niches in your base.

Then the user creates the labels that they care about- perhaps some variations of the population are meaningless? “lapsed customer, no pants, not cross-sold” could be potential to upsell, or perhaps they represent an audience that is truly uninterested in the pants-upsell promotion?

That’s a relatively simple example. Using feedback data, or more complex logic statements, truly shows the strength of these tools. Above, I have chosen binary yes/no choices, but you can use multiple answer logic, as well:
- Has purchased last 3 months, 6 months, 9 months, etc.
- Purchased above median, below, substandard, above standard, etc.

Creating the segmentation logic is good only as the analysis of those segments, determining things like the median price, pruning amounts on cells (according to email cost), and other factors. Make sure that the segmentation tool has appropriate and relevant reporting that will backup the thresholds of each segmentation filter.

I’m spurred to write this because a friend on Facebook- Jeremiah Owyang- attended a similar event, and I wish I’d gone:
Using Social Media to Grow Your Business.

What’s interesting to me is that Jeremiah chose to post the event and attendance on a social network. “Fishing where the fish are,” is how he explained it. There are so many layers of meaning here- that we examine the methods that we contact our customers, and use them ourselves in promoting our own efforts.

In a meeting with a client the other day about Facebook apps, he said, “How do we get them to come here,” the meaning being that merchandising and selling online as a dot-com retailer was the end goal. His bottom line inquiry made me think about how that wasn’t the goal, and yet it was the goal as well. Profit, and purchases are at the core of a retailer. From a CRM perspective, though, we look at the entire length of a customer’s interaction and relationship with the company. Brands, trust, and engagements that are not all around individual profit orders. It’s in between basic ordering and longterm relationships that social networks lie.

Little know-nothing startups like Scrabulous (play Scrabble on FaceBook) have become extremely well known because of an engaging Facebook application. As a consumer, I do “click over” to the Scrabble site. I have name recognition with the founders, I am aware of their new products, and various other fallouts from being a frequent user. Any company with online functionality and fun behavioral widgets understands the embracing of a large amount of new users like this, and how it can translate to customers down the line.

But back to Jeremy’s comment about fishing where the fish are- I see the social networks as a large customer base of people who have said: “I want to be contacted this way.” For those that remember, it’s the web portal business, but has finally matured. So yes, it’s about getting someone to click over- but it’s also getting someone interested and engaged in your product features, in a pond that they swim in.

Note: I want to write an upcoming post on the hilarity of Facebook’s opt-out system gone awry with Beacon. See: Mark Zuckerberg’s “Thoughts on Beacon”