Can you think of anything you did differently between your 80% efficiency batches and this one? IOW, the first step in troubleshooting is to ask what, if anything, changed. From Kia:
crush: If the crush is too coarse the pieces of endosperm will be too large for the gelatinization and enzymes to convert them at the chosen saccharification temperature and time. Evaluate the crush. Are there any uncrushed kernels? Are there any kernels that are cracked, but the endosperm doesn't come out easily? But keep in mind, that there is a limit in how much a tighter crush can improve conversion efficiency (especially if other mash parameters are off as well) and that it can be overdone. From my experience I suggest a crush done with a mill gap spacing between 0.5 and 0.8 mm (20 - 32 mil).
temperature: While any temperature below the upper temperature limit for the amylase enzymes (about 80C/175F) can lead to full conversion, only temperatures between ~63C/145F and 80C/175F are known to fully convert a mash with practical saccharification rest time. The exact temperature that should be chosen for the saccharification rest depends on the mash schedule, rest time, attenuation target and diastatic power of the malt. Check your thermometer against another one or at least against ice water (0C/32F) and boiling water (100C/212F; if you are close to see level).
pH: the enzymes, that are responsible for starch conversion, are not only sensitive to temperature but they are also sensitive to pH. The brewing literature and small scale mashing experiments (Limit of attenuation experiment) show that the mash pH should be between 5.3 and 5.7 (measured at room temperature) for optimal conversion efficiency. The mash pH is determined by the water/mash's mineral content and the gist composition. Check your waters residual alkalinity (RA) with John Palmer's residual alkalinity spreadsheet or, even better, check the mash pH. You won't need to buy a pH meter for this, using precision test strips like EMD's colorpHast strips are good enough.
diastatic power: this can be a problem with grists containing large amounts of highly kilned malts (e.g. dark Munich malt) or unmalted grains. Highly kilned malts contain fewer enzymes because the kilning process damaged more of them and the use of unmalted grains (e.g. rice or corn) dilutes the enzymatic power of the mash because they only add, if any, a small amount of enzymes to the mash. If the amount of enzymes in the mash is low, the mash may not be able to convert itself or the acceptable ranges for the other mash parameter got much smaller. A mash with low enzyme content may just need more time or a different mash schedule (step mash v.s. Single infusion mash) to be able to fully convert. If this is suspected to be the culprit, choose a different mash schedule or replace some of the enzymatic weaker malts with enzymatic stronger (e.g. Pale or Pilsner) malt.
time: the longer the mash rests while the enzymes are still active, the more of the starch can be converted. Once all starch has been converted time will not have an effect on the conversion efficiency anymore. But it may still have an effect on other wort properties like fermentability for example which is why we don't just stop mashing once the mash is converted. When mashing at lower temperatures or with enzymatic weak grists extending the saccharification rest length from 60min to 90min or longer may just do the trick for getting to complete conversion and satisfactory conversion efficiency.
water/grist ratio: the gelatinization and enzymatic activity require free water and in order to be fully completed, enough water needs to be available. This sets the lower limit of water/grist ratio for mashing. According to Briggs this lower limit is around 2 l/kg or 1 qt/lb [Briggs, 2004]. If too much water exists in the mash the enzymes may be to spread out to efficiently work on the starch. But that only happens at very thin mashes. It has been reported that even mashes as thin as 6 l/kg or 3 qts/lb work well. Based on this, it is unlikely that a mash was too thin for conversion.
mash schedule: Enzymatic weak mashes, mashes with undermodified malts (which are hard to find these days) or adjunct mashes may not fully convert when using the simple single infusion mash schedule. For such mashes it may be necessary to boil some of the grains/malt (decoction or cereal mashes) or use different saccharification rest temperatures. For enzymatic weak mashes it may be beneficial to add a dextrinization rest between 70C/158F and 75C/166F to speed up the alpha amylase and get the last bit of starch converted. But note that if there is a significant efficiency gain through this higher temperature rest it will be in the form of mostly unfermentable extract and thus raise the fermentability of the wort. This may need to be accounted for by using a lower temp saccharification rest.
dough balls: those can happen when the starch around a ball of malt gelatinizes and prevents water from reaching the malt inside that dough ball. If enzymatic activity is strong enough the gelatinized starch that is covering the ball will be converted and dissolved and eventually make way for the water to reach the dry malt inside. If that is not happening it causes conversion efficiency losses and even if the starch inside the dough balls is eventually freed, it may happen too late for full conversion of this starch. To avoid this problem make sure the mash is well stirred at dough-in. I also found that if the strike water temperature is above 60-65C (140-150F) adding grain to water reduces dough balls. At temperatures below this range barley starch doesn't gelatinize yet and dough balls are not likely to form.