Follow the yellow brick road to startup success

This is a guest post by Hubert, Palan, a good friend and classmate from Haas School of Business, UC Berkeley. Hubert (twitter: @hpalan) is the founder and CEO of ProductBoard.com, a platform for strategic product design and management headquartered in San Francisco, California. Prior to ProductBoard, Hubert was the Vice President of Product Management at GoodData, where he managed GoodData’s disruptive platform business, built the whole front-end product management team from the ground up and established and embodied modern principles of user experience designs.

An additional note – I see Hubert as the model for taking risks. He decided to launch a startup not because it was the only option available to him but when he was succeeding in his career and had multiple choices at his disposal.

Have you written your guest post yet?


yellow brick road
yellow brick road (Photo credit: hairchaser)

Let me tell you a short story. My wife, Jenna, and I went for a run early one morning in the Oakland hills. We had an idea where we wanted to go, but we didn’t have a map so we didn’t know how to get there. As we ran we asked several dog-walkers for directions along the way. Since it was after a rainy night, and I was running in very thin-soled running shoes, we asked if their recommended path would be muddy or not. As it turns out, different people have conflicting opinions about both the directions and the quality of the road. Eventually after a few wrong turns we found the right path, but of course contrary to what people said, it was pretty muddy.

Why am I telling you this? I recently quit my job at GoodData and started working on my own startup. Our morning run got me thinking about the challenges you have as a founder building a startup, or a new product. You have an idea of where you want to arrive – your dream target audience with a great need, that your perfect solution will satisfy.

You need to create a product roadmap that would navigate your team. Not only do you not have a map, you don’t even know if there are any roads out there. Muddy or otherwise.

So you head out and start asking for advice. You talk to advisors, investors and one potential customer after another trying to discover the roads and choose the best and shortest one. Advisors and investors give you conflicting advice about the best way, because even though they are great runners, they never ran quite the same route. Various potential customers suggest different features they would want, though they are not really sure if they even need them. They are like the dog-walkers, who are not sure about the route either, but since you ask, they make up an answer.

So you run in circles and take many wrong turns. You hope for smooth paths, and they turn out to be muddy. Hopefully though, you end up finding the right way and reaching your goal.

My friend and mentor Arthur J. Collingsworth always said: “Persistence, persistence, persistence.” So no matter if you are running in some hills, building a startup or working on a new product, persevere and keep on running.


Do you have to enchant your customers to gain loyalty?

What does it take to gain customer loyalty?
Beating their expectations is one way. But by how much?
Do you have to beat their expectations by a mile?
Do you have to forgo profits in the form of lower prices and higher service?
Can your business profitably beat customer expectations?For any marketer trying to gain customer loyalty in the form of repeat purchase, these are valid questions. After all there is no point in gaining loyalty of customers at the expense of profit.This article is about answering these questions using consumer behavior research.

Background and Hypotheses Development

Sometime back Tom Hulme sent me a tweet on his experience with Nespresso. Tom enjoyed using  his Nespresso machine but one day the water container broke. Tom said,

Did Nespresso price its part correctly?
Did it have to price it so low to gain loyalty?

I posited that Nespresso gave away too much, priced it incorrectly and should have given choices.

These discussions  led me to propose the  following two hypotheses

H1: Brands do not have to beat customer expectations by too much. They can get the sameeffect by beating it just enough.

H2: When customers are given choices at different price points, they will self-select themselves to the right version and will exhibit same loyalty as those receiving large price discount.

The loyalty here refers to attitudinal loyalty as there is no easy way to measure behavioral loyalty.

Experiment Design

I designed a between groups experiment to measure the difference between the stated attitudinal loyalty of different groups.  There are four groups in this experiment, all of them are filled in on Nespresso and were primed with a fixed willingness to pay of $30.

Since customers do not not what they are willing to pay and some of my experimental subjects may not know the cost of parts I used the price of $30 to normalize their willingness to pay.

Different groups were given different price rent by quoting them different price for the replacement part.

Group A:   WTP = $30, Quoted Price = $2.99
Group B:   WTP = $30, Quoted Price = $25.99
Group C:   WTP = $30, Quoted Price = $19.99
Group D:   WTP = $30,  Choices: Basic $9.99, Exact $19.99, Premium $28.99

Group B and Group C are similar but test different price points.

I designed the experiment using survey format (thanks to SurveyGizmo and its very powerful split testing functions) and ran it as a survey on people in my network and bunch of MBAs from Haas School of Business, Berkeley.

Respondents were asked to state their likelihood of repurchase  on a 6 point scale (a measure of loyalty). I also asked them to rate their likelihood to recommend the brand to others, more on this later.

Results

For testing the first hypotheses I compared the sample mean using 1 tailed t-test.  Between Group A ($2.99)  and Group B ($25.99) there was statistically significant difference (p=0.023) between the two samples. This could mean that beating customer expectation by a mile, in the form of very low price will have higher effect on loyalty than beating customer expectation just by a foot.

Between Group A ($2.99) and Group C ($19.99), the difference is not statistically significant (p =0.243). This is a critical finding. While $25.99 was no enough, $19.99 engendered the same level of loyalty as $2.99. That is a huge price difference. Brands do not have to give away the farm in  the name of loyalty. This also points to lost profit opportunity for Nespresso.

Next  let us take the second hypothesis that choices and self-selection (Group D) would perform at least as good as giving steepest price discount (the $2.99 option Group A).

Comparing sample means show there is statistically significant difference between mean likelihood ratings of Group A and Group D (p = 0.014). This is a big surprise for three reasons.

For one thing, when customers were given choices and self-select themselves to the version they prefer, they are more likely to feel ownership and increased utility.

Second, this Group was offered the same $19.99 price for the “Exact Match” version. This was the only option offered to Group C. While Group C showed no difference from Group A, this group did. Presence of choice negated any positive effect from $19.99 price.

Third,  if we looked at the sub-group  that chose the lowest priced Basic version ($9.99), there still is statistically significant difference between this sub-group and Group A.

One conclusion we can make is that presence of options for replacement parts causes customers to incur cognitive cost that is reflected in the form of low loyalty rating.  However, this requires further consideration before casting aside versioning.

One interesting corollary is the correlation between loyalty measured as intention to repurchase and likelihood to recommend. As I stated before, I asked respondents to rate both. There is very high correlation (0.99) between the two metric. Likelihood to recommend is not a better measure as contend.

Marketing Implications

Loyalty does not have to mean “delighting, enchanting, astonishing” customers. You can beat customer expectations by just enough. This is attitudinal loyalty and may not translate into behavioral loyalty. So in general using price discount to generate future sales is not recommended.

Statistical significance does not mean economic significance. The mean loyalty rating for lowest price group was 4.4 vs. 3.68 for $25.99 group. Will gaining loyalty at the cost of $22 per customer generate more profit in the form of future purchases?

For pricing replacement parts, brands need to do Willingness to Pay studies just as they do for the full product. There is no reason to sell the replacement part at cost due to fears of customer backlash. Same principles of value based pricing apply for parts.

While multi-version pricing is effective in most scenarios, offering choices for replacement parts comes at a cost to customer (See 4 costs of versioning). While versions enable profit maximization its effect on customer loyalty needs to be considered.