Show HN: Solving the problem of hiring salespeople | Hacker News Reader

Show HN: Solving the problem of hiring salespeople

idealcandidate.com

13 pointsbaudehlo9 comments

I can see the merits of this, but (scanning the front page) it looks like the primary tool for selecting candidates is a psychometric assessment.

First of all this is only data driven if we can show that candidates who score well are actually good sales people.

It is true that people with certain big four profiles for instance, tend to end up in sales. But are they more likely to be good salesmen? People with low intelligence or impulse inhibition tend to end up in prison, but are they effective criminals?

This is an exciting first step - by gathering quantitative data about the roles that the candidates go into, and how they perform, you could build the model you need. You probably don't have it yet though.

The other important thing is that different types of sales require very different people.

Sometimes, sales people need to penetrate a market where contacts are important, and deals are huge but take a long time to set up. They need to be the kind of person who can foster long term relationships, and maintain long term personal relationships with people and organizations.

Other times, deals are small, or relatively small, and what is important is volume. Maybe they need to rapidly create a shallow relationship, close the sale, and walk away. Maybe they need to cold call, and be especially tenacious in the face of rejection.

I'd be fascinated to see you create a questionnaire or similar that identifies the right person for that specific role.

What about administrating the tests to current employees, along with details from their bosses about performance. That way you could build a model of the kind of person who does well at that organization, at least generically.

The important thing about the way we calculate things here is that there isn't a "scoring well". You might max out the scores on certain groups of answers and it won't get you the job because we're comparing candidates to the top performers, who likely won't max out those profiles - having a more balanced personality could be far more important.

We have done some work pulling in Salesforce data directly so we can assess results - a feedback loop of sorts - once a candidate is hired we directly see the results of their performance. We aren't far enough along to publish results of that feedback loop though because all sales roles take a while ramping up.

  What about administrating the tests to current employees, along with details from their bosses about performance. That way you could build a model of the kind of person who does well at that organization, at least generically.

That's exactly what this does.

Even though I find psychometric test questions pretty cringeworthy in general, I do see the value of this service (anecdata on them actually working in a big corp environment: psychometric testing was part of a screening process which ultimately lead to me being offered a sales role I stayed at for 3 years and was rehired for after a career break, as opposed to the more junior one actually advertised, whose working conditions I certainly wouldn't have tolerated for more than a week). It's also slickly put together.

The negatives: I first questioned the value of the service when it limited the industries I could be interested in to five, especially before the evaluation process. A huge proportion of sales jobs hire more on personality match than anything else, and the overlapping categories ("accounting" is pretty strictly a subset of "professional services", and "advertising" of "media") don't help.

Even by the standards of psychometric tests the personality questions are repetitive. I certainly wouldn't have completed it without evidence there were interesting jobs the other side if it wasn't posted here; by the tenth question about whether I found it easy to read others' emotions I think even a really bad human recruiter would have taken the hint. Even accounting for the need to rephrase questions to check for consistency and some of those questions on basic empathy selecting for reliance on visual cues, there were too many. There aren't many salespeople who are going to honestly acknowledge an actual weakness in social skills in a system designed to rank their sales ability anyway (and similarly even serial liars like to create the impression they're generally truthful...except perhaps when asked about the lengths they'll go to for a sale).

On the other hand the sales questions, which seemed to be fewer in number and much better, seemed to miss some of the absolute classics in understanding salespeople's approaches like "salespeople should always be closing", "I'm best when focused on reaching as many new customers as possible", "understanding the customer is mostly about social skills".

For that matter, I'm still trying to work out why the eventual output scored me lowest for "analytical skills" (there's a joke in there somewhere...) but I think that reflects the balance of questions asked more than my actual relative weaknesses.

I actually think the idea has a lot of potential: there are vast differences in sales environments, often very poorly communicated by the actual ads, and usually vast numbers of salespeople with similarly disparate skillsets and homogenous CVs. But the questions could use some work.

Unrelated whinges: I don't particularly like or trust LinkedIn social signup even though I can understand why you'd use it (and contrary to popular belief there are good salespeople that don't use LinkedIn). I particularly don't like giving LinkedIn details to you and then having to create an imaginary North American address and zip code to continue the process, even though theoretically I could be willing to relocate a lot further than 150 miles, and many sales positions are remote...

I've been working on this for the last year now. The basic problem is that as a developer you have lots of ways to evaluate a candidate (coding tests). With hiring for Sales it's not that easy - the traditional way is to rely on an interview in which you go on gut instinct.

We're trying to change that by applying some science to the problem. The key idea is to find you sales people who match your current top performers' profiles using an assessment that is geared towards the sales process.

The system is built in Node.js, Angular and PostgreSQL, and uses some of the machine learning concepts I contributed to SpamAssassin.

Interesting concept. Why not change your tagline to "coding interviews for sales people" and specifically target the tech community. It also wasn't clear about "what's different".

IMO this is a huge problem in startup tech right now - having good sales abilities and also hiring people who do.

We're targeting tech companies right now purely because they tend to "get" this much better than other industries, but ultimately it's a much bigger market than just tech.

I'll pass on the feedback about the lack of "what's different" to the guy doing the site - it seems to be a bit of a pattern.

For the best test, check out Martin Seligman's work on using optimism as a proxy for sales success. He wrote about it in his book Learned Optimism.

He started this work when consulting for a large company that wanted to improve their sales hiring process.

Thanks - I've passed that on to our psychologist.

In general sales people are hired on a 12 month "understanding". The unwritten rule is they basically have 12 months to produce or get out.

The control is that they usually have lower base salaries than you'd think, but they can produce virtually unlimited income off commission if they can sell...usually with all kinds of thresholds in their employment contract on deals (e.g. 7% on deals <$50k 10% on deals >=$50k and <$250k).

It can take months to close a deal like that, so you can't just do a 90-day or GTFO trial like with other positions. But at the same time, if they aren't producing you'll know after a year. If you look at the resumes of professional sales persons, they'll usually have lots of ~12 month positions.

Good salespeople should also be bringing with them their contact lists and be able to start sourcing things very quickly. You have to make sure thought that their contacts are relevant to your company. A guy with deep insurance industry contacts might not be helpful in a B2C company.

They'll often also need to take over the pipeline that was in the works when they came on board, establish new contacts with that pipeline and go forward. They should also have basic managerial experience and know how to manage an inside sales person, a sales engineer, etc. They should be able to run a sales pipeline and figure out equitable commission sharing schemes with their team.

Something that's very hard to find it not only somebody who can do the above, but also actually sell. For example, if you sell complex software, finding a sales person who can provide even a basic demo of the software is like trying to skin a bag of rabid cats while you're in the bag. But you can use this problem as a filter. Give them access to your software and some information about it and have them come back in a week to do a 15-minute high-level sales demo and sell you your own software. If they can't at least do that, then you can pass and move on.

Reply on news.ycombinator.com