Multichannel marketing analytics best practice with Kevin Hillstrom

A liveblogging summary from ECMOD 2010

Online, Kevin is best known as @minethatdata, his moniker showing his passion for getting the most from your customer and sales data. I has worked at some of the biggest multichannel retailers in the world including Nordstrom, Eddie Bauer, and  Lands' End using an analytical driven approach to improve profitability.

I was pleased to be in the audience at the ECMOD Conference to learn from Kevin as he explained his consulting approach of multichannel forensics which he has applied over 50 companies in the last 3 years covering 1 billion purchase transactions!

Challenges of catalogue marketing

On a positive note  he revealed that on average c40% of customers should return over time to buy again, the Internet hasn't changed this greatly.

Instead, acquisition is the biggest challenge for majority of his clients - it's becoming harder to acquire new customers through time, e.g. mailing lists are now much less responsive.

Often asked by CEOs "Is Print dead"? His answer?

A subset of customers still highly responsive to catalogue marketing - these should be mailed more often, others less often or upweighted to email - what I call "Right-channeling" or "Right Touching". So identify those!

Kevin then ran through 35 strategies to consider. They have nice simple labels, so hope some are practical. Many of these targeting and testing approaches work for email marketing.

1. Consider smaller page count - effective for most clients.

2. Change contacts. Increase contacts for catalogue responsive, decrease number of contacts for others.

3. Order start. Use "order starter" offers upfront in catalogues or email marketing campaigns

4. Mail/phone responsive customers. Analyse and then flag as field in database.

5. Identify multichannels. customers who buy online after receiving catalogue. There is a continuum of customers who migrate to more online customer behaviour.

6. Identify online only shoppers. Can't matchback to catalogue, not-responsive to catalogues - again identify and create fields.

7. Rural shoppers. Catalogues tend to work 10% better with these (in the US at least)

8. Know your organics. These are customers who are brand loyal who are likely to buy anyway. E.G. One client is 13% another 50% another 87%. So can vary dramatically.

9. Do Mail/Holdout testing. Use an active testing strategy where 10% of active housefile does not receive catalogues or emails for 3 month period.

10. Use matchbacks. 100% accurate for understanding the impact of new customer acquisition.

11. Know email productivity. Not Opens or clicks - next to useless. Instead track beyond click to know % sales and again do Holdout testing. Determine true value of email revenue per email.

12. Merchandise + Email. A potential 20% upside when you can identify which types of merchandise work for the customer and upweight those.

13. Search and Print interaction. These work together. Catalogues tend to increase searches, so need to matchback search conversions to catalogue. Profit for search should get matchbacked to catalogue at the segment level.

14. Annual retention rate. Understand this and set targets, strategies to improve. If >60% loyalty programmes work particularly well to increase purchases/year. If < 30% you'd better be good at acquisition!

15. Business models. Catalogues are not essential - lot's of new options - have example of Bubbleroom.se who have 13,000 bloggers on their site offering 8% commission to these.

16. Merchandise path. Know path from first to third purchase and merchandising that affects these. But common behaviour doesn't exist for all.

17. Sales and promotions. 3 groups - will pay full-price, discount - only buy with promotion and mixed.

18. The iPad. A device tailor made for digital catalogues. "A fundamental shift in how customers buy". Buy one!

19. Indirect buying. Customers - understand shift between them.

20. Marketing = Humans. Human contact increases loyalty as you'd expect. So try to introduce cost-effectively.

21. Awareness. Social media rarely gives more than 2-4% of annual sales. But it does help generate awareness - acquisition potential is the greatest benefit of social media.

22. Content. Which content generates sales. Understand latency - number of visits between purchase.

23. Integrated data. Integrate web analytics, online sales through to store. One client follows 3 2 1 rule Visit 3 times a month, store twice and buys once.

24. Map the Future! Plot past channel performance to extrapolate to the future. Also forecast how customer behaviour changes.

25. Demographics. Probability of different demographics buying over different channels - phone, email, web. Correlate against merchandise preference, e.g. "newness" message.

26. Take risks. Measured risks of course!

27. Seasonality. Don't waste money mailing those who only buy at holiday period. Base on RFM analysis.

28. Capitalise on months 1-3. Well known that this is a "golden period".

29. Merchandise categories. Customer buy across categories are much more valuable than customers who are multichannel, so again identify, segment, target.

30. Small numbers. Micro-target. But often not practical to define/understand 256 segments, so his preference is a nice balance of 16 minimum!

31. Filtering. Give customer choice of channels - they will filter.

32. Retail shoppers. Behave differently.

33. Comparable segments. Benchmark performance through time E.g. two orders/year, last purchase 0-3 months ago.

34.  Clickstream data. "Sense and respond" Mail if customer visits site.

35. Digital profiles. Actionable segments, e.g. multichannel mavens, rotary phone, never buy online.

Made it - most of these are common sense but guess you'd need to buy the books, read the blog if you found these interesting.