The AI-driven world of surveys could use a personal touch.

Those who use ride sharing apps have doubtless noticed the ride is not the complete part of the story. Once you arrive, your app, somewhat intrusively, insists you complete a survey, rating the driver. The survey is Part Two of the journey. One must take care to answer it “correctly.” Inquiries follow if one doesn’t.

Which makes one wonder …

Deep learning (n, int): Techspeak for making something simple, often intuitively obvious, sound more sophisticated than it truly is. Faux profundity.

And ask this: Who gets to define “correctly?”

Further: What qualifies the person doing the defining?

“Thank you for choosing to enlist the expertise of the team at Abyss Technical Failure Analysis Services (ATFAS). If we could take a few moments of your time to complete the enclosed online service excellence survey, we’d be grateful for your candid feedback. We strive for continuous improvement, and your opinions are integral to that process. The following survey, activated by the link you clicked, will take only a few minutes. Many thanks for your business and, in advance, for your thoughtful answers.”

You’re welcome. Let’s move on.

“First question: Did you use our 1) X-ray Services; 2) CT Scanning Services; 3) Destructive Failure Analysis Services; 4) SEM/EDS/EDX Services; 5) Flying Probe Test Services; or 6) Other Services. Please check one. If your answer is 6, please explain.”

6

“Please explain.”

We used flying probe testing services, in combination with x-ray analysis, to nondestructively validate the connectivity and robustness of solder connections at certain critical device locations on our PCBA.

A series of long gray dashes appear at the bottom of the screen, resembling extended Morse Code, or even blank spaces of an application, waiting to be filled in, signed and submitted.

“Second question: On a 1-5 scale, with 1 being not helpful at all and 5 being extremely helpful, how helpful did you find the services you selected at ATFAS?”

3

“Please explain.”

You did a satisfactory job. You fulfilled our objectives and requirements. This is, in fact, what we paid you to do, and that is your job. Accordingly, you deserve neither condemnation nor praise. You simply performed to requirements, as you would be expected to do. Take our payment for services rendered and bank it.

The first, leftmost dash at the bottom of the page changes from gray to green.

“Please explain further. Why did you select the rating number 3?”

Was the prior answer insufficient? Also, it is our habit not to provide superior ratings, as there is always room for improvement in any service or activity open to subjective evaluation.

A second dash morphs from gray to green.

“Why did you not select ratings 4 or 5?”

You didn’t solve our problem. Your results were inconclusive. You did not pinpoint the source of failure at the I/O pins of the suspect BGA. Forgive the expression, but we were hoping/expecting to find a smoking gun, and you did not find one. Not entirely your fault, but expectations of success were high and went unfulfilled.

“What could we/should we do to deserve a higher rating for the services you used at ATFAS?”

Find our problem. That would rate a minimum of 4.

“What if our results are inconclusive, or we simply can’t find the source of your problem?”

Then the rating of 3 stands. It’s a results-driven world.

A caption suddenly emerges beneath the two green, formerly gray, dashes at the bottom of the screen. It says, “Good answer” also in green.

“Third question: On a 1-5 scale, with 1 being poor, and 5 being excellent or superior, how would you rate the level of customer care and service you received during your recent project with ATFAS?”

4

“Please explain.”

We rarely if ever give the uppermost available rating to anything. As previously noted, there is always room for improvement. It’s an imperfect world, in which many strive for perfection. Keep striving.

A third dash at the bottom of the page suddenly becomes green. Starting from the left margin, there are now three infiltrating green dashes, followed by a series of gray dashes.

“What could we/should we do in the area of customer service to enable you to raise your rating of ATFAS from 4 to 5?”

Cure cancer. Create permanent world peace. Eliminate poverty. Solve malnutrition and hunger. Tell us who really killed JFK. Little things.

“Please explain.”

I was being facetious. Probably nothing you can do to improve your rating at this point. You would have to provide tangible evidence of going significantly above and beyond customarily expected levels of service. I can’t provide examples, but like the Supreme Court justice who once defined pornography, I’ll know it when I see it.

“Please explain.”

I don’t know with certainty but will hazard some guesses. Answer emails promptly. Provide regular status reports ahead of my asking. Notify me of problems in process, their reasons, and your solutions to them. Explain the images I’m looking at on the screen to the satisfaction of colleagues who don’t understand what they’re looking at. Be patient with our crazy requests. Give us full value for the large amount of money we pay you (full value being defined and determined by us – again, we’ll know it when we see it).

“Please explain.”

I have nothing to add, other than to note that good service is something you should be providing anyway, and without the incentive and prodding contained in a customer survey. You collect money to do this, both in your revenue and in your salaries. This is not suburban Little League; there are no +9*trophies for participation. Get used to it. The technical term is maturity, also known in the vernacular as adulthood. Which leads to professionalism. Doing the right things when nobody’s looking. Nosy surveys don’t change that.

The algorithm relents, finally. A fourth dash turns green. The caption beneath the Morse Code dashes suddenly changes from “Good answer” to “Very good answer.” It’s happy. Almost home.

“Fourth question: On a 1-5 scale, with 1 being inferior and 5 being superior, please compare ATFAS as a supplier with similar suppliers or services you have used in the past. Does your recent experience with ATFAS lead you to conclude our services are better or worse than comparable services you have previously used?”

3. No better and no worse than comparable services.

“Please explain.”

We did exactly one order with you. Now you want our perspective on how your company fits the whole of human history, and commercial history to boot, after one order. Slow down, big fella. We’re just getting to know you.

“Please explain.”

We are basing our comparative rating of your company and its services on a sample size of exactly one. Accordingly, and because your service is new and mostly unfamiliar to us (other than this first order, which was fulfilled satisfactorily), we start you with what we believe is a fair rating, in the middle of the rating pack. You start from a neutral position, with nowhere to go but up, or down. Consider our rating as a challenge to do better. That’s the definition of constructive criticism. You did ask us to rate you, didn’t you? And now we must defend our choice of rating?

Just digest this: At this stage you are no better and no worse than other similar companies, with similar services performed. Work hard at earning our business, and adding to your sample size, and the ratings will take care of themselves and find their natural level.

A fifth dash immediately turns green. The machine is learning.

“Fifth question. Knowing what you know about ATFAS and its services, is there a service we lack that you need, or a shortcoming in our organization that you experienced, that in your opinion could use improvement? Please be as descriptive and as specific as you can.”

Provide imaging services for large objects in the sub-5 micron resolution range. The technical term is microfocus imaging. By large objects I mean server boards in the 500mm x 550mm range, weight 30-50lbs (including heat sinks), with 50-80 processors in BGA packages. Also stuffed with QFNs. Identify microcracks and their propagation on each device. In plain English, pick up small stuff very clearly. Find the failure nondestructively, and decisively, and large amounts of money are saved in diagnosis and repair. Everybody wins. Today that is an unfilled niche.

Also, better integrate your services with Industry 4.0 standards and capabilities.

“Please explain.”

Create the digital infrastructure to process large amounts of failure data (Exhibit A: the server board noted immediately above) and learn from it. Observe trends in the process of learning. Identify small problems from the learning before they grow into big problems. Pinpoint regressions from the mean before they violate control limits. Sentinel duty. Activate alarm bells when needed. Voilà: Failure analysis actually using big data and not simply warehousing it for marketing purposes.

A sixth dash turns green. The caption changes again, from “Good answer” to “Excellent answer.”

“Sixth and final question: Do you have any additional remarks, comments, commendations, or criticisms you’d like to add, that were not covered by the previous questions or your previous answers?”

Abyss Technical Failure Analysis Services’ services were adequate, to the extent that we have a second failure analysis project designated to send to you for inspection and evaluation. We hope you will apply the comments we’ve offered here in the spirit in which they were written, and this newest order will be subject to efforts at improved performance.

The entire dashed line at the bottom of the page turns green. And the caption now says, “Superior answer.”

Does this entitle us to a free replay? Where’s the happy face emoji? What have we won?

The algorithm is silent.

Words fail. •

Robert Boguski is president of Datest Corp. (datest.com); rboguski@datest.com. His column runs bimonthly.

• Prev
• Next