Revving up your recommendation engine

Yariella Coello, head of consultancy at data science and marketing services company Profusion, discusses how small businesses can use a recommendation engine.

We live in an age saturated with marketing messages and ever decreasing attention spans. Every business is fighting to break through to consumers. Marketing teams are spending an increasing amount of time trying to find the magic bullet that cuts through the noise and resonates with consumers. The old adage of targeting consumers at the right time, in the right way and with the right message still holds true. However, it is a mix that many businesses are still getting wrong.

Thanks to the sheer amount of data consumers produce every day, there simply is no excuse for poorly personalised and timed marketing. One way businesses can get their customers’ attention and loyalty is through recommendation engines.

Recommendation engines, otherwise known as personalisation engines, are a nifty data science tool that businesses can use to offer personalised product or service recommendations to customers based on previous purchases, browsing behaviour and other behavioural, demographic or declared data .

Using a recommendation engine, a business can predict what products a customer is likely to purchase in the future. This makes sending the right messaging at the right time a laughably easy task.

Recommendation engines are also not just reserved for large businesses. As long as you have enough sales and marketing data, any business can use one.

The data you use depends on your unique aims, your customers’ needs and what you have available. But by and large, everything from a customer’s online browsing behaviour and visits to your website, responses to previous marketing campaigns, loyalty card data and previous purchases can be used to fuel the recommendation engine.

If you want to get even more detailed, external data on weather, the economic environment and social media trends can be added to the engine to give more timely recommendations. Using this information, a recommendation engine could suggest items based on seasonal patterns or other buying behaviours influenced by economic, social and environmental factors.

Importantly, a recommendation engine moves beyond common sense suggestions (like recommending tennis balls to someone who has just bought a racket). Using the data you feed it, plus a constant feedback loop that tells the engine how well it’s recommending items, the engine could suggest items that the buyer had otherwise not considered but would certainly like – like sunscreen and a visor when there’s a sudden hot spell.

With the right input data, a recommendation engine can take into account an individual’s needs, wants and interests. The tennis player previously mentioned, may do the sport alongside regular workouts and long distance running. In which case, they may be interested in a foam roller to ease their aching muscles or supplements to aid recovery.

It’s also worth noting that when your recommendations are spot on, your customers’ loyalty to your business deepens. Giving good, relevant recommendations shows that you understand and know your customers. In the case of the tennis player, recommending items that they may not have known you stocked and which really helped their performance could make you their go-to shop for all their future sports equipment.

Recommendation engines are still relatively unknown or very much sit behind the scenes, although this is rapidly changing thanks to their high profile use by Amazon, Spotify and Netflix. It’s worth remembering, however, that you don’t have to have a large company to use a personalisation engine effectively – small business can also reap the benefits of the technology. Give it a few years and recommendation engines will be as commonplace as any car in the street. It’s worth investing in one now before the competition catches up.