IBM’s Deepak Advani talks innovation, Facebook and cognitive analytics

General manager of IBM Commerce Deepak Advani has spent most of his career leading various divisions of IBM, including most recently as GM of cloud and smart infrastructure, as well as spending time at Lenovo (which bought IBM’s PC business in 2005 and its server business in 2014) as CMO and senior vice president of ecommerce.

Advani was in Australia from the US last week to speak at the company’s customer engagement forum designed to educate and inspire marketers about the opportunities provided by its technology and software. Marketing sat down with him to discuss IBM’s new innovation program, ThinkLab, its partnership with Facebook, and the future of technology in marketing.


Marketing: You recently announced Facebook has joined your IBM ThinkLab innovation program. Can you give me a bit of background on how that concept works?

deepakDeepak Advani: We engage with many, many clients and some will say, ‘Hey, I have a problem that I need to solve and I can’t by software’, or there doesn’t seem to be an off-the-shelf solution to address my problem.

So IBM can you bring our best minds, researchers, data scientists, to spend a day spec-ing out how  we could solve the problem. It’s really client-centric innovation. So we mentioned this idea to Facebook and we said we’d love to partner with you.

Let me give you this example to make it a little more real:

analyse-theme-badgeWe had a large retailer that came to us and said that they’ve got their ecommerce business and then they’ve got their brick and mortar business, and they sell the same products on both but they run them as two independent businesses, it’s very common.

And they said in their holiday shopping season, they do their best job of planning assortments and their inventory but for the last several years what happens is they always run out of certain products on their website, and they always have excessive inventory in their brick and mortar stores. They they end up losing on both sides.

So they said, ‘IBM, can you help us with a problem like this?’ so we had our data scientists and researchers all work with the client and we came up with a pretty innovative solution, just for that one client, that said: any time an order comes in, we are going to route that order based upon optimising a bunch of different variables.

So what we’ll do is we’ll treat your inventory and your ecommerce inventory as a virtual pool. We re-routed the order, every order that came in, based on real-time analytics, on where is the best place to ship that from.

The first year we deployed it they had incurred $14 million of charges the year before, just for appeasements and expedited delivery. That went from $14 million to zero.

Then in the summer we productised, because we want to come up with new innovations that address specific problems clients have, then scale it to a lot of our other clients.

M: It also sounds clever from IBM’s point of view because it’s low-risk product development for you. So tell me about what you did with Facebook.  

DA: When we did did the Facebook partnership one of the things we said is that a lot of clients are struggling with attribution, you know, it takes eight to nine touches before somebody buys a product. Which touch had the most impact? So one of the things that Facebook wanted to show is if you included Facebook tactics as part of your overall marketing mix, what kind of lift did you get?

We said, well, this is another idea which we should recruit some clients to say, ‘Why don’t we do some studies where you run a controlled experiment of marketing mix without Facebook and you see what kind of return you get, and you do another one with Facebook and you see what kind of lift you get in that controlled environment?’. We talked to a lot of clients, we got over 30 that we worked with over the last six months, so we’ve got some real concrete feedback. And this is still happening everyday.

M: The other side to the partnership with Facebook is IBM Marketing Cloud using Facebook personal data to improve targeting and personalisation. Is this mostly around advertising within Facebook itself or are you also taking the next step to use the data to follow people to other sites as well?

DA: Facebook is very careful about the data that they have about their users, and they always keep that within Facebook, so there’s no personal data that ever goes from IBM cloud to Facebook cloud. The conversation we’re really having with Facebook is there’s certain users that we have, that we know, that have given us emails and opt-in permissions, you actually encrypt that email, and go to Facebook to say, ‘can we do something more relevant to them within the Facebook context?’ The second thing is the lookalike audiences where you can get 150,000 more people like that, so in that instance a lot of it is about more advertising within the Facebook world.

There are ongoing conversations with Facebook all the time going what else can we do together because we’re very aligned in one thing which is we want to help our clients serve their customers better.

M: It’s interesting to think about how Facebook fits into the marketing mix; it’s obviously very influential and they’ve got a lot of data. How do you see for marketers the importance of Facebook?

DA: I think Facebook is really important because marketers need to focus on where their customers are and their customers are spending a lot of time in Facebook. Part of the reason Facebook was interested in partnering with IBM by the way is because nine out of 10 retailers use IBM Commerce, eight out of 10 banks globally, nine out of 10 automotive companies… and we work with all of our clients on pulling together all of their data that they have in their transactional systems, a lot of them happen on IBM systems.

So you’ve got a lot of transactional data, a lot of our clients will have demographic data, from some data warehouse that they’ve used, they do a lot of surveys, there’s a lot of data that our clients have that’s not social data, and we’ve been helping those clients analyse the data, cleanse that data – because a lot of it is garbage in garbage out – if you don’t have a trusted platform around all that data, then you can’t do deep analytics on it, so you need data cleansing, data integration… so IBM worked with some of the largest clients on the planet, on how do you pull together all of the data that you have, and then Facebook comes with additional insights on all of your customers who are on social, and what clients want to do is really start to pull all of that together.

M: Speaking of the transactional systems, you mentioned on stage the importance of combining backend with front-end systems. What are your thoughts on marketing getting more involved with integrating everything?

DA: I think it’s really important; you could have the slickest marketing campaign, advertising campaign, but then your website goes down or you don’t have the thing in stock or it doesn’t arrive when they said it’s going to arrive – you’re done.

No one’s going to say, ‘I love that brand because their ads were so cute’. It’s about the experience, it’s about giving me what I value and then the experience. I read someplace that a promise broken anywhere is a promise broken everywhere. It’s about the weakest link because if you have that bad experience that’s what people walk away with.

M: And a challenge for marketers is to be taken seriously within companies.

DA: If we can just put on our futurist hats on and make some predictions, I really feel that marketing if they decide to stay in their own lanes and say, ‘I’m just going to do advertising and run campaigns’, then I think the importance of marketing will be less and less. What I envision and what I think is likely to happen is that marketing will increase in prominence in the organisation, because you talk to CEOs who will say their number one priority is better understanding customers.

They’re not going to ask finance to do that, they’re not going to ask supply chain to do that. So I think there’s an opportunity for marketing organisations and CMOs to align with the CEO to say, ‘Let me be the customer steward for you and let me then make sure that we’ve got enough metrics that we’re monitoring across the enterprise and across the overall experience’.

I think it will change over the next few years. I think CMOs will have a seat at the table and I think they’ll be the ones that are really carrying the flag for customer experience.

M: Good news for marketers. Tell me more about the future of (cognitive supercomputer) Watson. It was interesting that you said the future is about cognitive combining with predictive analytics; not taking over but complementing. Where do you see that headed?

DA: We are still very much in the early innings on cognitive, but already we’re seeing some really interesting things in that. For instance when they do predictive analytics, they look at historical data – so they’ll look at a sample of customers’ demographics, transactional history, behavioural information like they came to a website and what did they do, then you want to start looking at their attitudinal information and sometimes you’ll do market research to better understand that… So a lot of times when you bring all of these different types of data together and you run predictive algorithms, outcomes a model that can predict likelihood of a person’s response.  

With Watson, we’ve got new algorithms – things like personality insights and message resonance, that take your level of intelligence to another level.

With one of the clients who wanted to combine predictive and cognitive, we found 47% more results. They were looking for who is likely to accept coupons. If I start sending coupons to everybody, it costs me money to do it, it degrades my brand if people think, ‘Why are you giving me coupons I don’t use coupons’. Factoring in personality with predictive analytics they were able to more accurately identify those who value coupons and that’s who they offered coupons to.

So I think you’ll start to see use cases where marketing organisations will be able to better target the individuals, which is going to help you get higher return on investment, but I think you’re going to have relevant conversations with individuals who actually care.

How many of us still get spammed? It’s so irritating. I mean, when you see that coming from your favourite brands – come on, really? It degrades the brand and it annoys people, so a lot of these things will really help with that.


Check out Marketing’s coverage of Watson’s entrance into the industries of sport, fashion, education and customer service.


Michelle Herbison
BY Michelle Herbison ON 26 October 2015
Assistant editor, Marketing Magazine.