I accepted a job that I'm not really qualified for from a repeat client who I've enjoyed working with in the past. When they offered me the job, I told them that I wasn't qualified and offered to help them find someone who is. They encouraged me to learn on the job and said they would be patient, and encouraged me to accept, so I accepted reluctantly.

After me not being qualified, the second problem is that the job is probably impossible. The goal is to improve the profitability of the client's automated stock trading algorithm to a degree that I believe is not realistic. Before starting billing, I wrote a proposal for the client that clearly outlined the possibility that we would not achieve the goal. (But I don't think the client really accepted this, although they said they did, and due to my unqualified-ness, any statement I make asserting or rejecting the general feasibility of the project could be dismissed as not authoritative).

So I went in thinking I had my bases covered, but now I'm seeing that this was a mistake. Even with all my disclaimers, I should not have accepted a job I thought was highly likely to fail. Now I am continuing to do trial and error with no end in sight, and I am convinced that I am wasting my time and the client's time and money. I have not said as much to the client yet, that's what I need help with. How do I gracefully part with this job?

During planning, I broke the project up into four phases. The first two phases had deliverables that I achieved, however these deliverables alone are useless to the client. Now I am in the third phase, the deliverable for which is to continuing iterating until the client is either happy with the improvements or decides to stop the job for other reasons. I would be willing to refund the client for the time I've billed while in this phase. Should I offer? The contract itself specifies an hourly rate, but no provision for what happens if the goal can't be achieved. It also specifies no specific term of service.


5 Answers 5


This is a research and development effort. People funding research generally understand that the result may not be a 100% success, and goals are often just goals, or you may need to help your client understand this.

Your deliverables should include documentation on what didn't work, what you tried, anything fundamental you learned about the algorithm. It is valuable to any future efforts to rule out certain lines of effort. Generate a report of your efforts so far, laying out the approaches under consideration. Rate them by likelihood of working, status of investigation, etc.

Then have a meeting with the client with the goal of deciding where to go next. Don't apologize for not completing a deliverable that is basically "do research". One of the options might be stopping now, since you've already ruled out all the easy stuff, and everything else is extremely unlikely. Or maybe the client is completely happy to invest some additional money toward you spinning your wheels on the more uncertain stuff on the chance you get lucky.

  • 3
    Right. R&D is R&D and it may not yield maximal results.
    – Andrew
    Commented Aug 31, 2018 at 15:07

I think at this point you just sit down with the client and explain you have taken this as far as you can. You were honest about being uncomfortable with the job upfront, so this should not come as a complete surprise to your client.

Should I offer?

In this case it might be a good idea. Reiterate you have taken this one as far as you reasonably can, after giving it your best effort. See where the conversation with your client goes.

  • I think the context of your quote is important: "I would be willing to refund the client for the time I've billed while in this phase. Should I offer?" Are you recommending offering to refund for time already spent in Phase 3?
    – David K
    Commented Sep 4, 2018 at 13:34

The existing answers do a great job of explaining your contractual and ethical situation (that you are fine on both counts).

However, I think it's worth also adding some potential actions that can make a Research and Development situation less stressful for both you and the client.

At current, you've agreed to work through multiple phases with deliverables. However, the reality is this phase may never yield the deliverable you or the client wants. Continuing to research indefinitely can put a lot of stress on you, and potentially the client relationship. As such, I'd recommend the following:

  • Be upfront with the client, that you are uncertain of ever reaching a full-solution that meets their needs.

  • Agree that delivering logs and reports of where gains have/haven't been found, is a deliverable in itself.

  • Agree on more concrete timescales, where the client can re-evaluate whether to continue research or not.

Telling the client where you realistically are, is nothing out of the ordinary. On any project, if you felt it was no longer achievable - you would expect to tell them. In this case, the only difference is that you're willing to continue researching on the chance you find valuable results at some point.

By letting them know what you think of the prospects and getting an agreement from them that they wish you to continue researching - you can reduce your stress, and avoid worrying that you're not meeting their expectations.

For the deliverable itself, it's then important to agree what does and doesn't constitute value to the client. Again, you're simply managing expectations and ensuring that no matter what the outcome - neither of you are going to feel bitter that it didn't work out as hoped.

Logs, reports and analysis of what you've tried, where it may have led to something and why it didn't - are extremely valuable in research. If the client agrees (and it sounds like they will), making it clear that these will be delivered - regardless of a working product - ensures nobody gets left with nothing. This also helps reduce your worry that you're not providing real value - if the client agrees to these deliverables, you now have a minimum that you can deliver and can be satisfied you still did the job.

Finally, I'd recommend agreeing on more concrete timescales. As written in the question, it sounds like the client is happy to continue paying you in the hope you find something that works (whether you do or not); but obviously this will not continue forever. From personal experience, that can be extremely punishing - as you never know when to call it quits, and when the client will want to either.

By defining new timescales, where you will research and then evaluate the output/progress together - you give yourself and the client defined opportunities to call it quits without any hard feelings on either side.

Ideally, you want to have a set date where you will deliver the current progress and agree with the client whether to extend research for a further period - or finish up at that point. Again, this fixed date gives you certainty that you will deliver value at a defined point; and will avoid you feeling like you're exploiting or misleading them.

As others pointed out, research is an extremely valuable task and it's not unethical to be in this situation - where you feel you cannot deliver their goal. But to make the most of this situation - you need to turn what has become a fairly uncertain process, back into one with certainty, agreed expectations and that feels achievable to you.


As i see here, main issue is ethical.

Is that an option for you to wrap up the current state of phase 3 and suggest end of contract due to lack of progress and grim projection of continuation?

Would work well for this is to add some visual aids (graphs, tables, presentation) in order to get your point across.


You are unable, even with diligent investigation and research, to provide the algorithms needed. However, could you work with a domain expert and incorporate their solutions within your work? Not someone to finish the project in its entirety, but one to architect the solution which you then implement. Or,just a small piece that can be incorporated into future deliverables.

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .