Talk to the people behind the screen

Wil Reynolds once asked a colleague to give him a quick call at six o'clock at night. He only wanted to say thank you. He nearly gave her a heart attack.

It was her one year anniversary at Seer. He'd congratulated another staff member that morning and missed her, so he emailed asking her to call him. She wrote back asking if everything was okay. When she did call, he tells me at 24:08, her first words were, "oh my God, I thought I was being fired or something."

He'd met with her three or four times that year to tell her how much he loved her work. Didn't matter. A vague message from the boss after hours, and her brain filled in the worst. "That taught me something," Wil said. "People are taking clues from you and you're scary."

It's the best argument I know for picking up the phone, because text strips out the tone. Wil kept coming back to that idea in the last episode of the season.

1. 16

   Talk to the people behind the screen

   The fastest research tool is still a conversation. Pick up the phone.

   "The phrase that I try to use the most is 'audiences over algorithms'. Let's take time to just humanize the people behind the searches. Just talk to some people."

   Wil ReynoldsS01E19 Hear it 33:30

   > said hello to a human today. he said hello back. still processing.

What keyword research can't hear

Wil told me Seer had just started sitting people down in front of search results. "Click on these and tell me what you see, what you think about what you see."

What came back was stuff you would never learn doing keyword research. "You hear intonation, you hear frustration. You hear pleasure. Oh, this looks great." Then you ask why, and they tell you: because it's got this, and this, and this. "The whole time I'm like, oh man, I'm learning so much from this person."

A keyword tool gives you the words and a monthly number. It can't tell you the person typing them was annoyed, or relieved, or halfway through comparing three options and tired of all of them. Our most common keyword research mistake post is about the same gap: the language customers actually use, and the benefits they're after.

It also helps to know what they're looking at. We once analysed 2,000 keywords for our SERP formations study and found search results look very different depending on the kind of query. Watching someone react to one of those pages tells you which bits they trust and which they skip. No rank tracker will show you that.

Audiences over algorithms

The phrase Wil uses most is "audiences over algorithms", and the full interview on kwasi.com carries it in the title. He's been saying it for years. Dan Shure's Experts On The Wire episode with Wil from 2016 covers similar ground.

A lot of it is about measuring the right thing. Wil talks about reading psychology books because they help him "understand and empathize better with people and how people make decisions." Otherwise, he says, you're "working off of metrics like clicks instead of looking at people." Later in the episode he goes further: "We are not here to get them links or rankings." Rankings, he said, are just one of the steps to growing the client's business.

Everyone's having a day

Wil also told me about a psychology study he thinks about a lot. People watched a video of a man driving badly, cutting people off, and called him every name under the sun. Another group first saw him take a call from his wife about their baby, then watched the same driving. Their reaction flipped completely: "Oh my god, I hope things worked out."

"That study just went to show that the context, there's more going on than just the stimulus," he said. Then, honestly: "I try really hard to remember that. I'm not good at it."

Analytics is the stimulus. A bounce, a rage click, a half filled form. The call is where you find out about the baby.

Five conversations is plenty

The usual objection is time. You don't need much. Jakob Nielsen's long standing rule from usability testing is that "the best results come from testing no more than 5 users and running as many small tests as you can afford." His article explains why the returns drop off so quickly after that.

Who you talk to matters as much as how many. Near the end of the episode, at 44:51, Wil makes the case for teams spending time with people who aren't like them: "You build better products and solutions when the people who build it have broader perspectives." Five calls with your five happiest customers will confirm your persona. Five calls with people who nearly didn't buy will prove it wrong, which is more useful.

Your five calls

• Book five short calls. Twenty minutes each, with recent customers or people who enquired and didn't buy. Offer a gift card if it helps.
• Say why you're calling in the first line. Learn from Wil's anniversary call. "Nothing's wrong, I'd love your opinion on our website" saves everyone a fright.
• Show them a search. Share your screen with the results for your main keyword and borrow Wil's line: click on whatever you'd click and tell me what you think.
• Write down their exact words. Then compare them with your headings and your keyword list. The gaps are your next content plan.
• Bring someone along. Invite a designer or developer to one call. It's the quickest way to get everyone arguing about the customer instead of each other.

This was the last episode of season one, and Wil's version of digital empathy was the simplest one we got all season: "just talk to some people." Bobby, in the last few minutes, said he had "more questions now than when we began." Fair enough. Listen to the full episode, or read more about Wil and Seer Interactive.