Customers who bought this also bought… (or, Why can’t we all be Amazon?)
Jon Kelly, Wood for Trees Deputy MD, shares his thoughts on Next Best Action (NBA), marketing your product range and why charities can’t expect to emulate the online behemoths – yet. To learn more about Next Best Action, see our overview paper here. Jon will be speaking about NBA with our client Concern at the Institute of Fundraising National Convention at noon on Monday July 2nd 2018.
Here’s a story I heard in the very early days of Amazon’s recommendation engine. An avid fan of Manchester Utd. had just purchased a video of highlights of their Premier League-winning season. Just as this was delivered to his door he received an email from Amazon suggesting he may also be interested in owning a video of Arsenal’s title conquest the season before… Even the non-football fans amongst you I hope will get the irony of this.
This took place some 15 years ago now and machine learning algorithms have moved a long way since then. I’d hope for example that this season’s Manchester City fans aren’t going to be implored to view the exploits of Chelsea from last year. Nonetheless we can probably all recall receiving similarly off-beat suggestions based on our purchasing history.
Amazon does this kind of thing often, and in terms of their bottom line, I’m sure it works well for them. But I feel that it may not necessarily be for all of us.
Why? Well firstly, the Amazon model is built on having a wide and diverse set of products, and the ability to serve those products at any time. I have over the years had many lengthy discussions about what is or isn’t a product in charity terms. Is regular giving a product, or just a payment method? Is DRTV a product, or just a channel? Whichever way you cut it I think it is clear that most charities do not have Amazon’s breadth of product offering. Amazon don’t just offer you one thing – they give you a range of suggestions. If you only have a handful of products, then maybe offer them all?
The second issue for charities is the timing of these offers. This is partly cultural and partly down to the need for collateral, but charities tend to run their campaigns on a strict calendar schedule. Hence the “June Cash Appeal”, the “Autumn Upgrade Campaign”. There is no place in that calendar for sending Mrs Miggins a cash appeal in August just because that might suit her best.
That said, I’m not suggesting that we all simply stick with what we are doing – there are changes we can make. Supporters must be put at the heart of our fundraising strategies. We should not be led by organisational ‘needs’ and the inertia of the ‘we’ve always done it that way’ argument. There is an effective middle ground here, even if we can’t yet make it all the way to the Amazonian paradise.
As noted above, current practice is generally dictated by a calendar of appeals. These appeals are selected in chronological order and with a campaign-specific target for the volume of people in the campaign. This does not consider which campaign a supporter is most likely to respond to. In the most extreme cases, if there is a cap on communications sent to any given supporter, it can mean that they will be selected for campaign A and then omitted from campaign B even though they are much more likely to respond to campaign B. It can also mean supporters being included in campaigns that they have no likelihood of responding to, merely to fulfil quotas.
We have been working with our clients on a hybrid Next Best Action model, which maintains the idea of a campaign calendar – and all the benefits of that – but allows supporters to be prioritised across a group of campaigns based on their likelihood to respond. This helps to ensure that supporters have the most relevant communications for them picked from the calendar, and communications are sent to the supporters with the most interest in the current product or message.
To find out more,