Carsten’s blog

Realtime decision/offer management within the marketing master record

It started with a small idea. I wanted to have a marketing master record hosted directly within the BI platform where all analysis were made, predictive models and where there was access to all data. Usually marketing master records are outside of core data sources, e.g. billing, provisioning/network and similar platforms.

This idea was widely welcomed by the different business owners, but not by the business itself. Mostly due to past failed attempts in smaller projects. Even so, sometimes you just have to move forward and hope the business will follow. Challenge the existing business!

The marketing master record is hosted within SAS and to access it, I used (or actually simulated) something that SAS calls Treatments. Treatments are realtime recommendations. They are prioritised workflows that accesses data based on structured rules, that uses a UUID as input and then gives you back the Treatment ID’s for the ones that has a successful positive match.

When I initially started looking at Treatments, I unfortunately found Treatments wasn’t ideal for the purpose. SAS was inside a hosted environment and so was our CMS/EPiServer. This meant that the API that exposes the Treatment had to be able to handle realtime requests. Hence the fact that SAS is a BI platform, there’s no real way to guarantee a 99,99 % uptime. This is a heartbreaker for many, but when you know it, you just have to work around it.

You could either:

Build a proxy API that handles all requests
Export end results from all business logic to a flat-file

In reality there’s nothing that’s right or wrong… You only have to select the most feasible solution for your requirement.

One of the limitations of Treatments out-of-the-box is it uses a very strict approach without the ability to learn what actually works. This logic can of course be implemented. E.g. A/B testing with control groups as a self-learning service can be programmed directly into SAS, but then it wouldn’t be a out-of-the-box solution anymore. In the rarest of cases, you would actually need this. Just getting your product out there is the primarily goal right now.

If you want to use SAS as a marketing master record, you could easily go two steps beyond that, and use it for managing the actual content and not limit it to “what” has to be offered. That means that you could centralise content within the master record and push it to apps, social, web, self care, web banking, oss, etc. Really only limited to the integration points. I used it to personalise text versions, image-urls, etc. to be context aware. Originally there were basically only 2 versions, but very soon we found out having subsets was necessary.

Even though should maybe don’t want to invest in having inbound CRM, you should think about having a master record. Imagining if you knew what you already know!

SAS Institute and EPiServer

It’s been some time since I’ve last wrote something, but a lot has happened!

I’ve been in detailed looking into SAS Institute and EPIServer, and how to combine the strength and logic of SAS with EPi’s ability to deliver through its powerful CMS.

On SAS I’m looking into how Treatments are handled and how to expose personalized content in realtime on EPi. EPi is then used to build high quality online solutions with high user interaction giving not just relevant content and classic CRM, but actually a unique experience that takes CRM to the next level.

I will definitely bring a lot more detail on this as soon as I get a chance to sit at my MBA.

Predictive Commerce!

So!

Commerce is defined by Wikipedia as, “…defined as a component of business which includes all activities, functions and institutions involved in transferring goods from producers to consumers…”

The only issue when doing commerce is, that it’s pretty hard for a web shop to know your needs or when receiving a mail with an offer; what mood are you in, what is it exactly you’re looking for and what need is it the user actually wants fulfilled?

In the old days, when my mom and dad were my age, when they went to buy new clothes or any other retail store, the clerk would instantly know what was bought previously! If my mom asked for something, he would instantly know all history, what she really wanted and all sizes, color, etc. It’s pretty hard to duplicate the human interaction, but with big data, data driven communication, automated marketing, etc., basically means that a lot of data is being collected, used to engage and involve you as a user and automating the process making everything scaleable, while at the same time predicting your reactions with a greater probability!

Oh well…

What I really wanted to say was, that the new age of web shopping is coming and making things much more seamless, relevant and convenient!

What you should do is always keep track of who’s visiting your website. Even anonymous users; use whatever is needed, e.g. cookies to keep track of users. As soon as a user goes from anonymous to a known user, you would then have to connect all your data.

So what can you use that for?

When you get good a collecting data you will have to use it everywhere; the mails you send out, rearrange your website for when users accesses it and rearrange your offers based on impressions, clicks, etc. and you have to do it live. You should combine different metrics. So even though everything points in one direction, users most often don’t know what they want, but one thing is somewhat sure. Most users want what everyone else has. So take that into accountant and try to do cluster analysis and update your predictive algorithm on-the-fly!

It may seem a bit complicated for most, but it’s actually pretty easy and some modern platforms do it out-of-the-box. Personally I would suggest you have it tailored to fit your purpose. Like people, no companies nor web shops, are alike!

Relevant messages

## challenge ##

Mobile marketing is about promoting messages directly to leads and prospects. Being relevant and personal. Building a relationship with the user. This creates a degree of trust and builds loyalty, which we then can use to promote relevant offers towards qualified users.

## example ##

Mobile marketing through mobile can be done as easily as pushing todays offer in a message, Buy the new Chuck Taylor’s within the next 2 days and get it at half price!

The message itself has no flaws. A clear CTA at the beginning of the message (Buy). There’s no doubt what it is you’re selling (Chuck Taylor’s). The offer is time limited (within the next 2 days). Best of all the offer is pretty decent (half price).

So could it be done better?

Chuck Taylor’s are Converse’s designer shoes (in the classic Converse). This makes perfect sense for the 18-25 year olds, hipsters, with an income of 0 to 30.000 USD, living in the urban or surrounding communities and in college or entry-level professional and they are incredible tech-savvy on mobile devices!

So maybe it would have made sense to target the message towards those people?

## data driven communication ##

When using the information you have on your leads, prospects and customers, it’s called data driven communication. It basically means that you either target your message more precisely or only send it to a limited group of people; adding personalization in your message dramatically raises the conversion rate, e.g. Limited offer for everyone under 20. Get the new women’s Chuck Taylor’s within the next 2 days and get them for 35 USD!. In this example only 2 parameters were used to give a much more targeted offer.

## behavioral ##

Enrich your data with behavioral information; location, reactions to impressions or similar to specify the message even more, Limited offer for everyone under 20. Get the new women’s Chuck Taylor’s Washed Canvas at Converse on Broadway and get them for 35 USD

The message suddenly went from generic Chuck Taylor’s to a specific offer (women and offer under 20) and from that to a relevant offer (model and location).

## automated marketing ##

The reason why these types of offers can be somewhat complicated in the real life isn’t that the offers doesn’t work; if done right your target group is very small and with a small target group you maybe get a big conversion rate, but what that’s worth if the global revenue is limited?

The idea is to create a series of marketing messages and associate different criteria to each message and each time a message is met the message is automatically pushed to the user. In many situations, they user will experience as pull, why this is also a much more trustworthy way of communicating. If done right, the user will see this as an ongoing conversation that develops itself into promoting relevant offers.

So if you have your entire target group,

Women that are under 20 and on Manhattan,

Limited offer for everyone under 20. Get the new women’s Chuck Taylor’s Washed Canvas at Converse on Broadway and get them for 35 USD

Women under 20,

Limited offer for everyone under 20. Get the new women’s Chuck Taylor’s within the next 2 days and get them for 35 USD!

Everyone you don’t know,

Buy the new Chuck Taylor’s within the next 2 days and get it at half price!

The next big challenge you should think about is your community engagement, but more on that next time!

Gamification

I’ve recently started taking a class on gamification from the University of Pennsylvania through Coursera.org. Until now I really haven’t learned anything new, but then again… Everything seems much more clear now. I’ve taking a lot of NLP/coaching classes, management, economics and I’ve been working with loyalty programs over the last 1,5 years or so. Gamification takes all of these elements and puts them into meaningful boxes and kind of explains you how to use them and why you should do so!

Gamification that engaged me!

I quite fast realized that there’s many ways to do gamification and it isn’t a one size fits all. Take the Nike+ for instance. I would say that’s pretty much for beginners to intermediat runners, but it never got me to use it, and I actually bought everything needed for it. I also have a frequent flyers card for SAS – Scandinavian Airlines – but the only time I register my flights is when it’s done for me, and the same goes for all my retail memberships!

All these are of course gamifications in a more classic way. Like you buy something and you get points for it. Nike+ did some amazing stuff with their service, but I prefer just to run and take it easy without someone shouting in my ear!

Something that actually worked for me for a while was Foursquare. I hunted those points wanting to be the best (with an altime high of 482 points) and got those badges I wanted, but it was the badges and points I hunted and the purpose of Foursquare kind of faded and today I’m almost not using it at all!? The problem with Fourquare for me is that the social part – which they are trying to solve – isn’t really working. I check-in, write a small comment and/or take a picture while doing so and others can comment on it as well. The problem for me is – what’s the point? I know I’m a bit more social than average, with a Klout score of 60, but I already have Facebook and Twitter to be social on!? Facebook is for friends and Twitter is for news and a bit more serious tweets. When I used Foursquare the most it was with a friend of mine and we competed against eachother. That was the game part. The goal was points and badges and nothing else. When he didn’t really check-in anymore neither did I.

So Foursquare actually had a good gamification, but there wasn’t any real life purpose, I didn’t feel any progress (always status quo) and I quickly lost interest!

What would work for me?

I’ve been thinking a lot on the professors words and the things I’ve learned while being employeed at my current work! I would say that making progression is important so I don’t feel stuck in the middle and nothings happening. That’s eventually what made me drop out on video games! Then it has to make sense. I have to feel that whatever I’m doing makes sense for me and not just in the virtual world. I really don’t have any time for that. I don’t mind the virtual world (or in other words internet based technology/services) supporting me, but there has to be a bigger goal out there. When I ran a lot I did it for charity, like water in Africa, Diabetis, etc., and I felt good about “paying” for running. So that’s an example where I’m paying someone to participate and not the other way around (e.g. giving prizes or similar). And the most important thing is to compete against someone and I have to occationally win, and more frequent that I lose… I have to like always want to be the best and actually get those small pointless victories. If you look at people who really love what they are doing, the are really focused and into what they are doing; endorphins are pumping through their vains!

That is the kind of user experience I want. Or simply I want to be amazed, and of course I want everything what gamification typically has with the PBL!

Great! So where do we start?

Well start by finding out what is it you want people to do (besides spend money)? How can you turn that into a game where people freely would participate and have fun while they are doing it?

Remember how people will behave on different scenarioes. People can behave the opposite way in what you thought they would and how do you handle that? It’s like the economy. It’s been slowing down now for quite some time now, and what people do is that they spend less money because they feel that the future isn’t surtain. That’s actually the worst thing you could do and what happens it that the economy slows even further down and even more people lose their jobs because no ones buying what they are selling. So always think of how people actually would react; think about how you seriously would react!

So if I should be a bit more of a business man on this approach I would do a top down on the business idea on gamification, and then I would do a bottom up on the gamified business idea, and hopefully you have a match. It’s very important that the gamified business idea is the idea and not the business approach. If it isn’t, it will seem half hearted and will probably have a lot of lose ends!

Best traffic… SEO or Social?

What is SEO? I mean I know it’s search engine optimization. Oblivious. I just mean what is an actual optimization and what is it you want to achieve? More tragic doesn’t mean a better conversion rate. I actually means – in a traditional way – that you will get a lower conversion rate, but of course higher revenue since more people are buying.

Another way of doing SEO is through social awareness. Id you can create an awareness om each social medium and link them all together then you’re on the right way.

So where to begin?

Start by doing the medium thing. Think about what you samt to share om what medium. You don’t want to share the same on Facebook as on Twitter. When writing something always think in key words. Stat real and just don’t bull anyone on anything. Keep your promises and stay loyal. If you do those things, then your users will do the same. Remember that the most critic ones are the ones who care the most. Why would they else write?

Now you have to think about linking them together! Think about what it.is you want to communicate and do it directly on your own web. Then find the perspective for the communication for each medium and the do your posts with a link to the page/post. When it is indexed by a search engine, then the link to your post is also linked and it gets a higher rating.

Facebook is different. They have a closed eco system, but when someone likes something, then it’s posted on their wall for all friends (and often their friends). Your reach is a bit lower, but the value is much higher some the likes are comparable to word of mouth marketing. On Twitter you get both.

When using SEO try using an A-B approach to find what gives the best ROI. Remember that a true SEO strategy contains all channels and not just the search engines. When that’s said you would usually find relevant search words, taggings on your web and optimize the UX to fit how your site is indexed. Even small changes can give a big ROI. There are services on the web which can manage your ads/adwords, so the ones performing the worst will automatically be replaced by other ones. The advantage of this is that you very quick will get a feeling for what’s attracting your users.

Remember the fewer clicks, the higher conversion rate.

SEO will definitely get you a lot of traffic, but clicks aren’t everything if they generate nothing besides traffic. Social has a high value, but the organic growth can take time. I would suggest combining the both into your strategy. Another thing to think of is that social – if used the right way – is free where as SEO is CPC/CPM and that can easily go up.

Reach out, and just do it…

To get to your true reach of your message, you have to create a burning platform of your own. You have to spread your words across all media, and treat and manage each medium differently. You can’t post or blog a thousand words on Twitter nor it would be cheap to write too few words on your blog. You can’t just link your blog posts on Facebook, tweet links to your posts on Twitter. Your message has to be spread across the entire digital landscape and treat each medium as its own channel, with your message, formatted, designed, and massaged to have the greatest impact in that particular channel.

This is what cross channel marketing is!

Each time you turn to a medium, you have to ask why this medium and who’s the audience. What is it you want to say (that actually the easy part), but how the hard part… I think the advantage of using Facebook or Twitter is that those account are more or less for life. Once you have reached a “client” through a Page like, they are your customer untill you screw up. Imaging if you try to communicate through TV and how expencive that is, and you have to do this over and over again. It’s the fastest from a single mass communication point-of-view, but you have to think in that you would have to redo it again and again. When kicking of a new campaign, you could start by doing a TV commercial, but you should deffinately end by using some kind of cross channel CRM approach.

To get you started there are some basics you have to start with. You have to know what your audience/customers want. Else you can sell it to them regardless of the price. When you know that you have to get the word out there. No matter how got you are, “word of mounth” isn’t really good on the internet and your organic growth will be very slow. The best way is actually by simply adwording it, but adwords are expensive. So think about how you measure what works the best and optimize, and when I mean optimize, I mean like on a hourly basis. Think about that mobile ads work differnt from regular and with the right ads, your payoff is multiplied. There’s a bunch of services out there which automatically monitors your ad perfomance and manages them on-the-fly and not just a report as alot of the mainstream traditional services.

Good luck!

Conversational Marketing

What is conversational marketing to you? Some call it data driven marketing, others interactive, social, etc., but aren’t they all just the centering around the same subject?

It’s called conversational marketing because it goes in to a dialogue with the prospect on the product and interactively tries to convince that the product is the right one from a prospect point-of-view. Conventional marketing will try to convince you to “do this”.

Imagine that you browsed the web page of an mobile operator for the new iPhone. You go away again since you’re not sure. After a week or so you go in again. Price still too high so you subscribe to their Newsletter hoping for an eventual offer. Later you are logged into Facebook and sees an advert for a cheap subscription from the same operator you check it out, and browse between several until you find your favorite, but nothing really matched your criteria, but you did read a bit extra on one of them. Later you get your first newsletter and read about the new iPhone. You click it, but still sees the iPhone is still the same price. It is pricy. Now the conversational marketing kicks in. You clicked a link, which identifies you – still anonymous – but you’re now prospect A. The we page automatically rearranges the main offer on the page giving you the iPhone + the subscription you checked to a discounted price. You’re not sure yet. Seems too good. So you close the browser down. Immediately you get an mail saying that we saw you actually look at the offer, but didn’t go for it, so “… You didn’t accept our great offer. This offer was tailored to you. If you still want it, it will still be valid for the next two hours. Sure you don’t want it right away?“. You don’t buy it. Still a lot of money. After an hour you think oh well… You go to the web page. First thing you see directly on the main page saying “Welcome back… Click here to resume your offer weren’t to you!”

This is of course a simplified version. In real life you should try to combine at least text messages, mails, letters, web, customer care, stores, etc, so you have the entire cross channel thing going for you!

Remember to track it…

It’s always hard to say exactly how your campaign performed, but it’s always easy to celebrate that it looked great! This is a lot like any other project. You have to find out what it is you want to do, implement it and then find out how to track it. It’s obvious much harder to find out how to track it after the campaign has finished.

Before you start…

Start by identifying your KPI. Remember any KPI has to be measurable or else it isn’t a KPI – it’s just something else.
When you have found your KPI’s, you have to find out what it is you want to know? Let’s say you can track how many sign-up for a product, but is that what you want to report on or is more financial numbers like what was the conversion rate, ARPU/CLV, etc. and even though you know those things, is it good or bad?
Set up control groups. If you don’t have anything to benchmark up against, it’s more or less pointless to make any kind of performance measurements. Then it’s much more a question of just having surveillance and reporting on KPI’s, but not really making any conclusions
You have to find out when you think it makes sense to do the measurement. Remember if you do the same campaign over and over again, e.g. everyrone who signs up for a product, then make sure that you performance the same measurements from a time perspective. E.g. if you want to know how many has clicked through from a mail to your web site, then you should wait around 5 days and see. The next time you pull your numbers, you have to only look at the people who clicked from day 0 to day 5 and not just anyone who has clicks. If you do any other way you will end up having numbers that are greater the longer you wait and that just not fair.

Connecting the dots…

When you carefully have designed your campaign. Think about how you want to implement it. Have you thought the entire process through? Everything is a process! Everything starts with a process. So start by using only a pen and a piece of paper. Draw a box for each thing which happens and call each box KPI #1, KPI #2, etc., and when you have drawn up each box in the process don’t do anything before you know how to track it. If everything is tracked, then you can always afterwards find out what it is you want to use it for (besides the obvious).